Claude Code Skills Marketplace

Professional Claude Code skills marketplace for development, research, content, and residential floor plan searches.

---

🌟 Essential Skill: skill-creator

⭐ Start here if you want to create your own skills!

The skill-creator is the meta-skill that enables you to build, validate, and package your own Claude Code skills. It's the most important tool in this marketplace because it empowers you to extend Claude Code with your own specialized workflows.

Why This skill-creator?

This is a production-hardened fork of Anthropic's official skill-creator, born from building real skills and hitting every wall the official version doesn't warn you about.

The official skill-creator tells you what to build. Ours also tells you what not to try — and why.

You're trying to...	Official	This Fork
Research before building	"Check available MCPs"	Ordered search protocol with decision matrix: Adopt / Extend / Build
Create a skill interactively	Prose-based instructions	Structured AskUserQuestion checkpoints — user never loses context
Avoid common mistakes	No guidance	Cache edit warnings, prerequisite checks, security scan gate
Know the architecture options	Not mentioned	Inline vs Fork decision guide with examples (choosing wrong silently breaks your skill)
Validate before shipping	Basic YAML check	Expanded structural validator plus provenance-checked old-vs-new capability audit; packaging re-verifies the completed review instead of trusting a marker
Match validation cost to risk	One full create-test-review loop	Risk tier and evaluation spend are separate: targeted checks first; paired benchmarks require explicit authorization or a decision-bearing evidence plan plus opt-in
Catch security issues	No tooling	security_scan.py with gitleaks integration — hard gate before packaging
Learn from real failures	No failure cases	Battle-tested methodology with documented failure patterns and gotchas
Distill past conversations safely	Not covered	Explicit local manifest, message-level time window, redaction, opaque source IDs, ignored .enrich/ staging, and manual promotion into references/scripts
Turn approved artifacts into skill behavior	Not covered	Artifact-corpus distillation workflow: script-measured pattern extraction (≥3-artifact threshold), cataloging≠distillation gate, invariant-vs-register boundary, independent completeness audit
Survive concurrent sessions on one skill repo	Not covered	git-ref baselines, re-read-before-write, HEAD check before commit, own-paths-only staging, one-bump-per-outcome versioning
Ground knowledge skills in evidence	General advice	Authority ladder from real calls and machine-readable specs through production code, plus executable-example smoke checks and evidence-boundary rules
Have both installed at once	Coin flip — the two descriptions are near-identical	Detects the clash on trigger and offers a one-command, reversible SessionStart routing hook (only ever installed when both coexist); the official plugin stays usable by explicit request
Your own skill collides with an installed plugin	Not covered	generate_supersede_kit.py stamps the same conditional routing kit into your skill, plus a measured precedence decision guide (rename → description tiebreaker → hook → disable)

Full methodology: skill-creator/references/skill-development-methodology.md

Quick Install

In Claude Code (in-app):

/plugin marketplace add daymade/claude-code-skills

Then:

1. Select Browse and install plugins
2. Select daymade/claude-code-skills
3. Select daymade-skill (includes skill-creator)
4. Select Install now

From your terminal (CLI):

claude plugin marketplace add https://github.com/daymade/claude-code-skills
# Marketplace name: daymade-skills (from marketplace.json)
claude plugin install daymade-skill@daymade-skills

What You Can Do

After installing skill-creator, simply ask Claude Code:

"Create a new skill called my-awesome-skill in ~/my-skills"

"Validate my skill at ~/my-skills/my-awesome-skill"

"Package my skill at ~/my-skills/my-awesome-skill for distribution"

Claude Code, with skill-creator loaded, will guide you through the entire skill creation process - from understanding your requirements to packaging the final skill.

📚 Full documentation: daymade-skill/skill-creator/SKILL.md

Live Demos

📝 Initialize New Skill

✅ Validate Skill Structure

📦 Package Skill for Distribution

---

🚀 Quick Installation

Install Inside Claude Code (In-App)

/plugin marketplace add daymade/claude-code-skills

Then:

1. Select Browse and install plugins
2. Select daymade/claude-code-skills
3. Select the plugin you want
4. Select Install now

Generate Installation Commands

These scripts print commands for you to run in your terminal. Download and run the script locally to choose from the featured-plugin menu or make a custom selection.

macOS/Linux:

bash -o pipefail -c 'curl -fsSL https://raw.githubusercontent.com/daymade/claude-code-skills/main/scripts/install.sh | bash'

Windows (PowerShell):

iwr -useb https://raw.githubusercontent.com/daymade/claude-code-skills/main/scripts/install.ps1 | iex

Manual Installation

Add the marketplace:

claude plugin marketplace add https://github.com/daymade/claude-code-skills

Marketplace name is daymade-skills (from marketplace.json). Use @daymade-skills when installing plugins. Do not use the repo path as a marketplace name (e.g. @daymade/claude-code-skills will fail). In Claude Code, use /plugin ... slash commands. In your terminal, use claude plugin ....

Essential Skill (recommended first install):

# skill-creator ships inside the daymade-skill suite
claude plugin install daymade-skill@daymade-skills

Documentation Suite (shared namespace for document workflows):

claude plugin install daymade-docs@daymade-skills

This suite exposes related skills under one namespace, including:

/daymade-docs:docs-router
/daymade-docs:doc-to-markdown
/daymade-docs:mermaid-tools
/daymade-docs:pdf-creator
/daymade-docs:ppt-creator
/daymade-docs:docs-cleaner
/daymade-docs:excel-automation
/daymade-docs:read-docx-review

These skills ship as a bundle — there are no separate single-skill plugins. All documentation skills live under daymade-docs/ and install together from the suite. The router handles automatic selection. The specialist commands remain available for manual use; For the retained PPT command, read the compatibility notice.

Apple Platform Suite (shared namespace for macOS and iOS development/operations):

claude plugin install daymade-macos@daymade-skills

/daymade-macos:capture-screen
/daymade-macos:developing-ios-apps
/daymade-macos:macos-cleaner
/daymade-macos:macos-load-doctor
/daymade-macos:macos-permissions
/daymade-macos:macos-watchdog

These skills are bundle-only under daymade-macos.

Codex Suite (shared namespace for Codex workstation setup, assisted coding, and visual exploration):

claude plugin install daymade-codex@daymade-skills

Codex CLI and Desktop users can install the same suite through Codex's plugin marketplace:

codex plugin marketplace add daymade/claude-code-skills
codex plugin add daymade-codex@daymade-skills

/daymade-codex:codex-image-gallery
/daymade-codex:local-codex
/daymade-codex:design-style-picker
/daymade-codex:interaction-design-board
/daymade-codex:codex-1m-context-window-setup

These skills are bundle-only under daymade-codex.

Claude Code Operations Suite (shared namespace for Claude Code power-user workflows):

claude plugin install daymade-claude-code@daymade-skills

This suite bundles the skills that extend Claude Code itself — cross-project prior-work retrieval across code, docs, Skills, meetings, WeChat archives, and conversation history; fast local conversation discovery across Claude Code and Codex; session recovery; CLAUDE.md tuning; version-synced Lark CLI routing; troubleshooting; statusline configuration; export repair; marketplace development and suite consolidation; terminal screenshot rendering; usage analysis; and multi-provider model switching:

/daymade-claude-code:claude-code-ops-router
/daymade-claude-code:local-conversation-history
/daymade-claude-code:read-claude-code-history
/daymade-claude-code:read-codex-history
/daymade-claude-code:continue-claude-code-work
/daymade-claude-code:continue-codex-work
/daymade-claude-code:claude-skills-troubleshooting
/daymade-claude-code:claude-md-progressive-disclosurer
/daymade-claude-code:statusline-generator
/daymade-claude-code:fixing-claude-export-conversations
/daymade-claude-code:marketplace-dev
/daymade-claude-code:terminal-screenshot
/daymade-claude-code:claude-usage-analyst
/daymade-claude-code:claude-switch-models-setup
/daymade-claude-code:read-claude-web-conversation
/daymade-claude-code:claude-migrate-memory-to-doc
/daymade-claude-code:claude-code-hooks
/daymade-claude-code:prior-work-retrieval
/daymade-claude-code:lark-cli-router
/daymade-claude-code:claude-code-ping-start-5h-quota
/daymade-claude-code:agent-web-search-setup
/daymade-claude-code:tech-selection

Installed names render as daymade-claude-code:<skill> under a single shared namespace. These skills are bundle-only — install the suite to get all members.

Financial Data Suite (shared namespace for investment-research and financial data workflows):

claude plugin install daymade-financial@daymade-skills

This suite bundles the skills that fetch and analyze financial data — Bigdata.com (RavenPack) structured financials and sentiment, US equity fundamentals via yfinance, Gangtise (岗底斯) OpenAPI research suite orchestration, A-share news and policy aggregation, A-share pharmaceutical sector daily reporting, structured devil's-advocate pressure-testing of investment theses, and adversarial due diligence on inflated benchmark claims:

/daymade-financial:financial-router
/daymade-financial:bigdata-skill
/daymade-financial:financial-data-collector
/daymade-financial:gangtise-copilot
/daymade-financial:ashare-news-fetcher
/daymade-financial:daymade-sector-research
/daymade-financial:pharma-daily-report
/daymade-financial:devils-advocate
/daymade-financial:benchmark-due-diligence

Installed names render as daymade-financial:<skill> under a single shared namespace. These skills are bundle-only — install the suite to get all members.

Install Other Skills:

# Research-driven improvement of a specified existing Skill
claude plugin install skill-evolver@daymade-skills

# GitHub operations
claude plugin install github-ops@daymade-skills

# Teams communication
claude plugin install teams-channel-post-writer@daymade-skills

# Local Claude/Codex agent messaging
claude plugin install peer-message@daymade-skills

# Ghostty terminal session snapshot/restore
claude plugin install ghostty-use@daymade-skills

# Repomix extraction
claude plugin install repomix-unmixer@daymade-skills

# AI/LLM icons
claude plugin install llm-icon-finder@daymade-skills

# CLI demo generation
claude plugin install cli-demo-generator@daymade-skills

# Cloudflare diagnostics
claude plugin install cloudflare-troubleshooting@daymade-skills

# UI design system extraction
claude plugin install ui-designer@daymade-skills

# YouTube video/audio downloading
claude plugin install youtube-downloader@daymade-skills

# Secure repomix packaging
claude plugin install repomix-safe-mixer@daymade-skills

# Full audio suite (ASR + transcript correction + meeting minutes + TTS)
claude plugin install daymade-audio@daymade-skills


# Video comparison and quality analysis
claude plugin install video-comparer@daymade-skills

# QA testing infrastructure with autonomous execution
claude plugin install qa-expert@daymade-skills

# Prompt optimization using EARS methodology
claude plugin install prompt-optimizer@daymade-skills

# CCPM skill registry search and management
claude plugin install daymade-skill@daymade-skills

# Promptfoo LLM evaluation framework
claude plugin install promptfoo-evaluation@daymade-skills

# Twitter/X content fetching
claude plugin install twitter-reader@daymade-skills

# Skill quality review and improvement
claude plugin install daymade-skill@daymade-skills

# GitHub contribution strategy
claude plugin install github-contributor@daymade-skills

# Windows Remote Desktop / AVD connection diagnosis
claude plugin install windows-remote-desktop-connection-doctor@daymade-skills

# Product analysis and optimization
claude plugin install product-analysis@daymade-skills

# Scrapling CLI extraction and troubleshooting
claude plugin install scrapling-skill@daymade-skills

# Tencent IMA knowledge base companion and installer
claude plugin install ima-copilot@daymade-skills

# Export Douban (豆瓣) book/movie/music/game collections to CSV
claude plugin install douban-skill@daymade-skills

# Terraform operational traps and multi-environment reliability patterns
claude plugin install terraform-skill@daymade-skills

# Evaluate any LLM endpoint across speed, concurrency, protocol, and quality
claude plugin install llm-eval-harness@daymade-skills

# Video/GIF memes with motion-tracked image overlays
claude plugin install meme-creator@daymade-skills

Standalone plugins can be installed independently; suite members install together with their suite.

---

🇨🇳 Chinese User Guide

For Chinese users: We highly recommend using CC-Switch to manage Claude Code API provider configurations.

CC-Switch enables you to:

• ✅ Quickly switch between different API providers (DeepSeek, Qwen, GLM, etc.)
• ✅ Test endpoint response times to find the fastest provider
• ✅ Manage MCP server configurations
• ✅ Auto-backup and import/export settings
• ✅ Cross-platform support (Windows, macOS, Linux)

Setup: Download from Releases, install, add your API configs, and switch via UI or system tray.

Complete Chinese Documentation

For full documentation in Chinese, see README.zh-CN.md.

---

📦 Other Available Skills

github-ops - GitHub Operations Suite

GitHub operations through gh CLI and GitHub APIs, with explicit targets, impact previews, and independent state verification.

When to use:

• Creating, viewing, or managing pull requests
• Managing issues and repository settings
• Managing collaborators, teams, organization permissions, and 2FA enforcement
• Querying GitHub API endpoints
• Working with GitHub Actions workflows
• Automating GitHub operations

Key features:

• Verified mutation workflow for PRs, issues, repositories, and Actions
• Parallel and superseded PR convergence
• Organization access, member privileges, and 2FA preflight
• REST, GraphQL, and documented UI-path selection
• Enterprise GitHub support

🎬 Live Demo

---

docs-router - Daymade Document Routing

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:docs-router)

Selects the bundled specialist for document conversion, Word and PDF production, Mermaid images, macOS Excel automation, scanned PDFs, DOCX review extraction, or documentation cleanup. It reads the selected specialist's full instructions and required references. Specialist slash commands remain available for manual use. For the retained PPT command, read the compatibility notice.

---

doc-to-markdown - Document Conversion Suite

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:doc-to-markdown)

Converts documents to markdown with Windows/WSL path handling and PDF image extraction.

When to use:

• Converting .doc/.docx/PDF/PPTX to markdown
• Converting saved HTML/HTM with source-link checks and explicit heading offsets
• Extracting images from PDF files
• Processing Confluence exports
• Handling Windows/WSL path conversions

Key features:

• Multi-format document conversion
• PDF image extraction using PyMuPDF
• Windows/WSL path automation
• Confluence export processing
• Helper scripts for path conversion and image extraction

---

mermaid-tools - Diagram Generation

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:mermaid-tools)

Extracts Mermaid diagrams from markdown and generates high-quality PNG images.

When to use:

• Converting Mermaid diagrams to PNG
• Extracting diagrams from markdown files
• Processing documentation with embedded diagrams
• Creating presentation-ready visuals

Key features:

• Automatic diagram extraction
• High-resolution PNG generation
• Smart sizing based on diagram type
• Customizable dimensions and scaling
• WSL2 Chrome/Puppeteer support

🎬 Live Demo

---

claude-code-ops-router - Claude Code Setup Router

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-code-ops-router)

Routes plugin and Skill repair, marketplace work, statusline, model profiles and source sync, 1M context-window repair, usage and quota timers, memory migration, and exported .txt repair to one bundled specialist. The selected Skill's full instructions are read at use time; its original slash command remains available for manual use.

---

statusline-generator - Statusline Customization

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:statusline-generator)

Configures Claude Code statuslines with multi-line layouts and cost tracking.

When to use:

• Customizing Claude Code statusline
• Adding cost tracking (session/daily)
• Displaying git status
• Multi-line layouts for narrow screens
• Color customization

Key features:

• Multi-line statusline layouts
• ccusage cost integration
• Git branch status indicators
• Customizable colors
• Portrait screen optimization

🎬 Live Demo

---

teams-channel-post-writer - Teams Communication

Creates educational Teams channel posts for internal knowledge sharing.

When to use:

• Writing Teams posts about features
• Sharing Claude Code best practices
• Documenting lessons learned
• Creating internal announcements
• Teaching effective prompting patterns

Key features:

• Post templates with proven structure
• Writing guidelines for quality content
• "Normal vs Better" example patterns
• Emphasis on underlying principles
• Ready-to-use markdown templates

🎬 Live Demo

---

peer-message - Local Claude/Codex Agent Communication

Follow the local coordination workflow for task/resource declarations, native preflight installation, recovery and receiver checks.

An experimental paired-network route adds invitations and cited answers from owner-selected documents through existing Claude/Codex hosts. It requires an operator-provided relay; the public hosted network and user-adoption claims are not part of this preview.

Install: claude plugin install peer-message@daymade-skills

Follow the routing contract for native communication and transport gaps.

When to use:

• Sending a dependency, pause, handoff, or completion notice to an independently identified target outside the current native tools' scope
• Finding replies to a specific coordination message without manually inspecting local message stores
• Reaching a Claude inbox from a script or another product when native tools do not cover it; a third-party provider alone is not a reason to use the fallback
• Unblocking messages held for per-message manual approval on an unattended endpoint (crossSessionInbound)
• An inbound peer message asserting facts about your session or shared state, or asking you to pause/release — verify the premise against its own authority before acting
• Another session's uncommitted edits, lock, or branch is in your way on a shared checkout — verify it is live, then ask the owner before waiting or working around it
• Broadcasting one explicit coordination message to a reviewed target list

📚 Execution entry and detailed SOPs.

---

repomix-unmixer - Repository Extraction

Extracts files from repomix-packed repositories and restores directory structures.

When to use:

• Unmixing repomix output files
• Extracting packed repositories
• Restoring file structures
• Reviewing repomix content
• Converting repomix to usable files

Key features:

• Multi-format support (XML, Markdown, JSON)
• Auto-format detection
• Directory structure preservation
• UTF-8 encoding support
• Comprehensive validation workflows

🎬 Live Demo

---

llm-icon-finder - AI/LLM Brand Icon Finder

Access 100+ AI model and LLM provider brand icons from lobe-icons library.

When to use:

• Finding brand icons for AI models/providers
• Downloading logos for Claude, GPT, Gemini, etc.
• Getting icons in multiple formats (SVG/PNG/WEBP)
• Building AI tool documentation
• Creating presentations about LLMs

Key features:

• 100+ AI/LLM model icons
• Multiple format support (SVG, PNG, WEBP)
• URL generation for direct access
• Local download capabilities
• Searchable icon catalog

🎬 Live Demo

---

cli-demo-generator - CLI Demo Generation

Generate professional animated CLI demos and terminal recordings with VHS automation.

When to use:

• Creating demos for documentation
• Recording terminal workflows as GIFs
• Generating animated tutorials
• Batch-generating multiple demos
• Showcasing CLI tools

Key features:

• Automated demo generation from command lists
• Batch processing with YAML/JSON configs
• Interactive recording with asciinema
• Smart timing based on command complexity
• Multiple output formats (GIF, MP4, WebM)
• VHS tape file templates

🎬 Live Demo

---

cloudflare-troubleshooting - Cloudflare Diagnostics

Investigate and resolve Cloudflare configuration issues using API-driven evidence gathering.

When to use:

• Site shows ERR_TOO_MANY_REDIRECTS
• SSL/TLS configuration errors
• DNS resolution problems
• Email Routing aliases, destination verification and forwarding delivery
• Cloudflare-related issues

Key features:

• Evidence-based investigation methodology
• Comprehensive Cloudflare API reference
• SSL/TLS mode troubleshooting (Flexible, Full, Strict)
• DNS, cache, and firewall diagnostics
• Agentic approach with optional helper scripts

---

ui-designer - UI Design System Extractor

Extract design systems from reference UI images and generate implementation-ready design prompts.

When to use:

• Have UI screenshots/mockups to analyze
• Need to extract color palettes, typography, spacing
• Building MVP UI matching reference aesthetics
• Creating consistent design systems
• Generating multiple UI variations

Key features:

• Systematic design system extraction from images
• Color palette, typography, component analysis
• Interactive MVP PRD generation
• Template-driven workflow (design system → PRD → implementation prompt)
• Multi-variation UI generation (3 mobile, 2 web)
• React + Tailwind CSS + Lucide icons

🎬 Live Demo

---

ppt-creator - Compatibility Notice

Read the retirement and installation-compatibility notice before using the legacy daymade-docs:ppt-creator command.

---

youtube-downloader - YouTube Video & Audio Downloader

Download YouTube videos and audio using yt-dlp with robust error handling and automatic workarounds for common issues.

For spoken-content text, reuse existing transcripts first, then fetch original subtitles; download audio for ASR only when those checked routes miss. File-download requests keep the existing workflow.

When to use:

• Downloading YouTube videos or playlists
• Extracting audio from YouTube videos as MP3
• Experiencing yt-dlp download failures or nsig extraction errors
• Need help with format selection or quality options
• Working with YouTube content in regions with access restrictions

Key features:

• Native yt-dlp clients first; optional PO Token setup after an explicit token diagnostic
• Browser-cookie verification for “not a bot” prompts (privacy-friendly)
• Audio-only download with MP3 conversion
• Format listing and custom format selection
• Output directory customization
• Proxy-aware downloads for restricted environments

🎬 Live Demo

---

repomix-safe-mixer - Secure Repomix Packaging

Safely package codebases with repomix by automatically detecting and removing hardcoded credentials before packing.

When to use:

• Packaging code with repomix for distribution or sharing
• Creating reference packages from proprietary codebases
• Security concerns about accidentally exposing credentials
• Pre-commit security checks for hardcoded secrets
• Auditing codebases for credential exposure

Key features:

• Detects 20+ credential patterns (AWS, Supabase, Stripe, OpenAI, etc.)
• Scan → Report → Pack workflow with automatic blocking
• Standalone security scanner for pre-commit hooks
• Environment variable replacement guidance
• JSON output for CI/CD integration
• Exclude patterns for false positive handling

🎬 Live Demo

Coming soon

---

transcript-fixer - ASR Transcription Correction

Install: claude plugin install daymade-audio@daymade-skills (suite-only — invoked as daymade-audio:transcript-fixer)

Correct speech-to-text (ASR/STT) errors with a Stage 1 dictionary pre-filter, a required Native AI whole-transcript pass, and an audio-backed human gate for unresolved wording, names, numbers, and entities.

When to use:

• Correcting meeting notes, lecture recordings, or interview transcripts
• Fixing Chinese/English homophone errors and technical terminology
• Reviewing one exact transcript with its original audio instead of searching a global queue
• Improving future transcripts without turning one-off fixes into broad dictionary rules
• Collaborating with teams on shared correction knowledge bases

Key features:

• Stage 1 + Native AI correction pipeline; Stage 1 alone is never completion
• Frozen review packets and checked file/segment coverage for split or resumed Native reviews; malformed results stay unready and valid work can be reused
• Exact-file review queue, deep-linked dashboard, timestamped audio playback, and machine-readable zero-pending readback
• Conservative pattern learning: file-only, dictionary, roster, and context have separate admission rules
• Domain-specific dictionaries (general, embodied_ai, finance, medical)
• SQLite-based correction repository
• Team collaboration with import/export
• GLM API integration for AI corrections
• Cost optimization through dictionary promotion

Example workflow:

# Initialize and add corrections
uv run scripts/fix_transcription.py --init
uv run scripts/fix_transcription.py --add "错误词" "正确词" --domain general

# Run full correction pipeline
uv run scripts/fix_transcription.py --input meeting.md --stage 3

# Open only this transcript's unresolved items and listen before deciding
uv run scripts/review-dashboard/server.py --file "/absolute/meeting.md"

# Read the human verdicts back; pending_total must be 0 before final delivery
uv run scripts/fix_transcription.py \
  --list-review --review-file "/absolute/meeting.md" \
  --review-status all --json

📚 Documentation: See daymade-audio/transcript-fixer/references/ for workflow guides, SQL queries, troubleshooting, best practices, team collaboration, and API setup.

Requirements: Python 3.10+ and uv. Native AI correction uses the active agent; an external API key is needed only for the optional agent-less API route.

---

video-comparer - Video Comparison and Quality Analysis

Compare two videos and generate interactive HTML reports with quality metrics and frame-by-frame visual comparisons.

When to use:

• Comparing original and compressed videos
• Analyzing video compression quality and efficiency
• Evaluating codec performance or bitrate reduction impact
• Assessing before/after compression results
• Quality analysis for video encoding workflows

Key features:

• Quality metrics calculation (PSNR, SSIM)
• Frame-by-frame visual comparison with three viewing modes:
  • Slider mode: Drag to reveal differences
  • Side-by-side mode: Simultaneous display
  • Grid mode: Compact 2-column layout
• Video metadata extraction (codec, resolution, bitrate, duration, file size)
• Self-contained HTML reports (no server required, works offline)
• Security features (path validation, resource limits, timeout controls)
• Multi-platform FFmpeg support (macOS, Linux, Windows)

Example usage:

# Basic comparison
python3 scripts/compare.py original.mp4 compressed.mp4

# Custom output and frame interval
python3 scripts/compare.py original.mp4 compressed.mp4 -o report.html --interval 10

# Batch processing
for original in originals/*.mp4; do
    compressed="compressed/$(basename "$original")"
    output="reports/$(basename "$original" .mp4).html"
    python3 scripts/compare.py "$original" "$compressed" -o "$output"
done

🎬 Live Demo

Coming soon

📚 Documentation: See video-comparer/references/ for quality metrics interpretation, FFmpeg commands, and configuration options.

Requirements: Python 3.8+, FFmpeg/FFprobe (install via brew install ffmpeg, apt install ffmpeg, or winget install ffmpeg)

---

qa-expert - QA Test Planning, Execution & Reporting

Set up test cases, execution tracking, bug reports and QA handoffs using AAA test structure and OWASP security categories.

When to use:

• Setting up QA infrastructure for new or existing projects
• Writing standardized test cases following Google Testing Standards (AAA pattern)
• Implementing security testing (OWASP Top 10 coverage)
• Executing comprehensive test plans with automatic progress tracking
• Filing bugs with proper P0-P4 severity classification
• Calculating quality metrics and assessing project release criteria
• Enabling autonomous LLM-driven test execution
• Preparing QA documentation for third-party team handoffs

Key features:

• One-command initialization: QA directories, templates and example CSVs
• Autonomous execution: Follow the approved project plan with result tracking and bug reports
• Runtime compatibility: Compare adapter outputs with the target runtime's actual semantics and distinguish modeled inputs from native execution
• Release acceptance: Apply the project's release policy; an overall pass rate does not override a failed critical user task
• Ground Truth Principle: Prevents doc/CSV sync issues (test docs = authoritative source)
• Bug tracking: P0-P4 classification with detailed repro steps and environment info
• Onboarding: Guide for new QA engineers
• LLM prompts: Ready-to-use prompts for specific QA tasks
• Metrics report: Test execution progress, pass rate, linked bugs and built-in check results

Example usage:

# Resolve qa-expert-dir from the loaded Skill; use a fresh project root
uv run "<qa-expert-dir>/scripts/init_qa_project.py" my-app "<project-root>"

# Report metrics after test execution
uv run "<qa-expert-dir>/scripts/calculate_metrics.py" "<project-root>/tests/docs/templates/TEST-EXECUTION-TRACKING.csv"

Initialization creates templates. Before execution, replace example rows, supply test specifications and verify the project harness. Follow the execution SOP for preparation, recovery and handoff. The metrics command reports results; it does not decide release acceptance.

📚 Documentation: Start with qa-expert for test selection, execution and reporting.

Requirements: Python 3.8+, uv

---

prompt-optimizer - Prompt Engineering with EARS Methodology

Transform vague prompts into precise, well-structured specifications using EARS (Easy Approach to Requirements Syntax) - a methodology created by Rolls-Royce for converting natural language into testable requirements.

Methodology inspired by: 阿星AI工作室 (A-Xing AI Studio), which pioneered combining EARS with domain theory grounding for practical prompt enhancement.

When to use:

• Converting loose requirements into structured specifications
• Optimizing prompts for AI code generation or content creation
• Breaking down vague feature requests into atomic, testable statements
• Adding domain theory grounding to technical requirements
• Transforming "build X" requests into detailed implementation specs
• Learning prompt engineering best practices with proven frameworks

Key features:

• EARS transformation: 5 sentence patterns (ubiquitous, event-driven, state-driven, conditional, unwanted behavior)
• 6-step optimization workflow: Analyze → Transform → Identify theories → Extract examples → Enhance → Present
• Domain theory catalog: 40+ frameworks mapped to 10 domains (productivity, UX, gamification, learning, e-commerce, security)
• Structured prompt framework: Role/Skills/Workflows/Examples/Formats template
• Advanced techniques: Multi-stakeholder requirements, non-functional specs, complex conditional logic
• Complete examples: Procrastination app, e-commerce product page, learning dashboard, password reset
• Theory grounding: GTD, BJ Fogg Behavior Model, Gestalt Principles, AIDA, Zero Trust, and more
• Progressive disclosure: Bundled references (ears_syntax.md, domain_theories.md, examples.md)

Example usage:

# Before (vague)
"Build me a password reset feature"

# After EARS transformation (7 atomic requirements)
1. When user clicks "Forgot Password", the system shall display email input field
2. When user submits valid email, the system shall send password reset link valid for 1 hour
3. When user clicks reset link, the system shall verify token has not expired
4. When token is valid, the system shall display password creation form requiring minimum 12 characters, 1 uppercase, 1 number, 1 special character
5. When user submits new password meeting requirements, the system shall hash password with bcrypt and invalidate reset token
6. When user attempts password reset more than 3 times in 1 hour, the system shall block further attempts for 1 hour
7. If reset token has expired, the system shall display error message and option to request new link

# Enhanced with domain theories
- Zero Trust Architecture (verify at each step)
- Defense in Depth (rate limiting + token expiration + password complexity)
- Progressive Disclosure (multi-step UX flow)

# Full prompt includes Role, Skills, Workflows, Examples, Formats

🎬 Live Demo

Coming soon

📚 Documentation: See prompt-optimizer/references/ for:

• ears_syntax.md - Complete EARS patterns and transformation rules
• domain_theories.md - 40+ theories mapped to domains with selection guidance
• examples.md - Full transformation examples with before/after comparisons

💡 Innovation: EARS methodology eliminates ambiguity by forcing explicit conditions, triggers, and measurable criteria. Combined with domain theory grounding (GTD, BJ Fogg, Gestalt, etc.), it transforms "build a todo app" into a complete specification with behavioral psychology principles, UX best practices, and concrete test cases - enabling test-driven development from day one.

---

local-conversation-history - Cross-Provider History Entry Point

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:local-conversation-history)

See the Skill's usage and routing contract for history recall, interrupted work and conversation-backup health checks.

---

read-claude-code-history - Read Local Claude Code History

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:read-claude-code-history)

Inspect prior work without resuming the old task. Start with the result router for the current workflow, prerequisites and coverage limits.

Choose the requested result:

• Unknown Session or remembered topic: indexed discovery
• Known Session: chronological evidence
• Original command and its receipt: tool call/result evidence
• Verbatim human wording: human-input export
• Deleted or overwritten files: recovery workflow

Use the commands in those entries rather than a separate README recipe.

---

docs-cleaner - Documentation Consolidation

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:docs-cleaner)

Consolidate redundant documentation while preserving all valuable content.

When to use:

• Cleaning up documentation bloat across projects
• Merging redundant docs covering the same topics
• Reducing documentation sprawl after rapid development
• Consolidating multiple files into authoritative sources

Key features:

• Content preservation: Never lose valuable information during cleanup
• Redundancy detection: Identify overlapping documentation
• Smart merging: Combine related docs while maintaining structure
• Validation: Ensure consolidated docs are complete and accurate

🎬 Live Demo

Coming soon

---

skills-search - Local Repositories and CCPM Search

Find Skills in configurable local repositories first, then approved sources and the CCPM registry. Preserve existing CCPM installation and management commands.

When to use:

• Finding skills for specific tasks (e.g., "find PDF skills")
• Installing skills by name
• Listing currently installed skills
• Getting detailed information about a skill
• Managing your Claude Code skill collection

Key features:

• Local-first discovery: Persist repository paths and trust tiers outside the Skill package; inspect every configured owned source before expanding
• Progressive reading: Search cached name/description metadata, then open the selected Skill at its exact Git commit
• Registry search: Search CCPM registry with ccpm search <query>
• Skill installation: Install skills with ccpm install <skill-name>
• Version support: Install specific versions with @version syntax
• Bundle installation: Install pre-configured skill bundles (web-dev, content-creation, developer-tools)
• Multiple formats: Supports registry names, GitHub owner/repo, and full URLs
• Skill info: Get detailed skill information with ccpm info <skill-name>

Example usage:

# Search configured local repositories first (<skill-dir> is this loaded Skill)
uv run --script <skill-dir>/scripts/local_sources.py search pdf --tier owned

# Expand to the external registry when needed
ccpm search pdf              # Find PDF-related skills
ccpm search "code review"    # Find code review skills

# Install skills
ccpm install skill-creator                # From registry
ccpm install daymade/skill-creator        # From GitHub
ccpm install skill-creator@1.0.0          # Specific version

# List and manage
ccpm list                    # List installed skills
ccpm info skill-creator      # Get skill details
ccpm uninstall pdf-processor # Remove a skill

# Install bundles
ccpm install-bundle web-dev  # Install web development skills bundle

🎬 Live Demo

Coming soon

📚 Documentation: See daymade-skill/skills-search/SKILL.md for configuration, coverage and commands

Requirements: Python 3.10+, Git and uv for local search; CCPM CLI only for external registry/install/manage operations

---

pdf-creator - PDF Creation with Chinese Font Support

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:pdf-creator)

Create professional PDF documents from markdown with proper Chinese typography using WeasyPrint.

When to use:

• Converting markdown to PDF for sharing or printing
• Generating formal documents (legal filings, reports)
• Ensuring correct Chinese font rendering

Key features:

• pandoc + WeasyPrint conversion pipeline (dual backend: WeasyPrint or headless Chrome)
• Built-in Chinese/Japanese/Korean (CJK) font fallbacks with auto CJK code-block rendering
• Theme system (default for formal docs, cjk-auto for content-driven tables, warm-terra for training materials, mobile for phone reading)
• A4 layout defaults with print-friendly margins
• Batch conversion scripts

Example usage:

uv run --with weasyprint scripts/md_to_pdf.py input.md output.pdf

🎬 Live Demo

Coming soon

📚 Documentation: See daymade-docs/pdf-creator/SKILL.md for setup and workflow details.

Requirements: Python 3.8+, pandoc (system install), weasyprint (or Chrome as fallback backend)

---

claude-md-progressive-disclosurer - CLAUDE.md Optimization

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-md-progressive-disclosurer)

Optimize user CLAUDE.md files using progressive disclosure to reduce context bloat while preserving critical rules.

When to use:

• CLAUDE.md is too long or repetitive
• Need to move detailed procedures into references
• Want to extract reusable workflows into skills

Key features:

• Section classification (keep/move/extract/remove)
• Before/after line-count reporting
• Reference file and pointer formats
• Best-practice optimization workflow

Example usage:

"Optimize my ~/.claude/CLAUDE.md using progressive disclosure and propose a plan."

🎬 Live Demo

Coming soon

📚 Documentation: See claude-md-progressive-disclosurer/SKILL.md.

---

promptfoo-evaluation - Promptfoo LLM Evaluation

Configure and run LLM evaluations with Promptfoo for prompt testing and model comparisons.

When to use:

• Setting up prompt tests and eval configs
• Comparing LLM outputs across providers
• Adding custom assertions or LLM-as-judge grading

Key features:

• promptfooconfig.yaml templates
• Python custom assertions
• llm-rubric scoring guidance
• Built-in preview (echo provider) workflows

Example usage:

npx promptfoo@latest init
npx promptfoo@latest eval
npx promptfoo@latest view

🎬 Live Demo

Coming soon

📚 Documentation: See promptfoo-evaluation/references/promptfoo_api.md.

Requirements: Node.js (Promptfoo via npx promptfoo@latest)

---

developing-ios-apps - iOS App Development

Install: claude plugin install daymade-macos@daymade-skills (suite-only — invoked as daymade-macos:developing-ios-apps)

Build, configure, and debug iOS apps with XcodeGen, SwiftUI, and Swift Package Manager.

When to use:

• Setting up XcodeGen project.yml
• Fixing SPM dependency or embed issues
• Handling code signing and device deployment errors
• Debugging camera/AVFoundation problems

Key features:

• XcodeGen project templates
• SPM dynamic framework embedding fixes
• Code signing and provisioning guidance
• Device deployment and troubleshooting checklists

Example usage:

xcodegen generate
xcodebuild -destination 'platform=iOS Simulator,name=iPhone 17' build

🎬 Live Demo

Coming soon

📚 Documentation: See developing-ios-apps/references/xcodegen-full.md.

Requirements: macOS + Xcode, XcodeGen

---

twitter-reader - Twitter/X Content Fetching

Fetch Twitter/X post and article content. Single-post text goes through the fxtwitter mirror API (login-free, key-free, full long-form body); X Articles with images go through fetch_article.py; the Jina API path remains as a fallback with a known intermittent-ban risk.

When to use:

• Retrieving tweet content for analysis or documentation
• Fetching X Articles (long-form) with all images downloaded locally
• Extracting images and media from posts
• Batch downloading multiple tweets for reference

Key features:

• Single posts: fxtwitter mirror — no key, no login, full tweet.text body including note_tweet long-form
• No JavaScript rendering or browser automation needed
• No Twitter authentication required
• Returns markdown-formatted content with metadata
• Supports both individual and batch fetching
• Includes author, timestamp, post text, images, and replies
• Environment variable configuration for secure API key management

Example usage:

# Single post text (preferred): fxtwitter mirror, no key needed
curl -sS "https://api.fxtwitter.com/USER/status/TWEET_ID" \
  | python3 -c "import json,sys; t=json.load(sys.stdin)['tweet']; print(t['text'])"

# X Article with images (full mode)
uv run --with pyyaml python scripts/fetch_article.py \
  "https://x.com/USER/status/TWEET_ID" ./Clippings

# Jina fallback (intermittent — see SKILL.md risk note)
export JINA_API_KEY="your_api_key_here"
curl "https://r.jina.ai/https://x.com/USER/status/TWEET_ID" \
  -H "Authorization: Bearer ${JINA_API_KEY}"

# Batch fetch multiple tweets
scripts/fetch_tweets.sh \
  "https://x.com/user/status/123" \
  "https://x.com/user/status/456"

# Fetch to file using Python script
python scripts/fetch_tweet.py https://x.com/user/status/123 output.md

🎬 Live Demo

Coming soon

📚 Documentation: See twitter-reader/SKILL.md for full details and URL format support.

Requirements:

• curl (pre-installed on most systems) — single-post path needs nothing else
• Python 3.6+ (for Python scripts)
• Jina.ai API key (only for the Jina fallback; intermittent-ban risk, and the shared repo key is currently out of balance / 402)

---

macos-cleaner - Intelligent macOS Disk Space Recovery

Install: claude plugin install daymade-macos@daymade-skills (suite-only — invoked as daymade-macos:macos-cleaner)

📚 Documentation: macOS Cleaner; for uv cache diagnosis or retirement, enter the Skill's uv target semantics.

Example usage:

My Mac is running out of space, help me analyze what's using storage

---

fact-checker - Document Fact-Checking

Verify factual claims in documents using web search and official sources, then propose corrections with user confirmation.

When to use:

• Fact-checking documents for accuracy
• Verifying AI model specifications and technical documentation
• Updating outdated information in documents
• Validating statistical claims and benchmarks
• Checking API capabilities and version numbers

Key features:

• Web search integration with authoritative sources
• AI model specification verification
• Technical documentation accuracy checks
• Statistical data validation
• Automated correction reports with user confirmation
• Supports general factual statements and technical claims

Example usage:

# Install the skill
claude plugin install fact-checker@daymade-skills

# Fact-check a document
"Please fact-check this section about AI model capabilities"

# Verify technical specs
"Check if these Claude model specifications are still accurate"

# Update outdated info
"Verify and update the version numbers in this documentation"

🎬 Live Demo

Coming soon

📚 Documentation: See fact-checker/SKILL.md for full workflow and claim types.

Requirements:

• Web search access (via Claude Code)

---

skill-reviewer - Skill Quality Review & Improvement

Review and improve Claude Code skills against official best practices with three powerful modes.

When to use:

• Validating your own skills before publishing
• Evaluating others' skill repositories
• Contributing improvements to open-source skills via auto-PR
• Ensuring skills follow marketplace standards

Key features:

• Self-review mode: Run the bundled reviewer backed by canonical skill-creator validation
• External review mode: Clone, analyze, and generate improvement reports
• Auto-PR mode: Fork → improve → submit PR with additive-only changes
• Evaluation checklist: Frontmatter, instructions, resources verification
• Additive-only principle: Never delete files when contributing to others
• PR guidelines: Tone recommendations and professional templates
• Reliable automation: Distinguishes review findings from invocation/runtime failures with structured JSON output
• Evidence-based design review: Score six dimensions with exact source citations; keep documented design quality separate from measured task benefit
• Read-only batch review: Inventory collections against declared host contracts, prepare complete selected-source packets, and export JSON/CSV/Markdown with explicit coverage and invalid decisions quarantined

Example usage:

# Install the skill
claude plugin install daymade-skill@daymade-skills

# Self-review your skill
"Validate my skill at ~/my-skills/my-awesome-skill"

# Review external skill repository
"Review the skills at https://github.com/user/skill-repo"

# Auto-PR improvements
"Fork, improve, and submit PR for https://github.com/user/skill-repo"

🎬 Live Demo

Coming soon

📚 Documentation: See daymade-skill/skill-reviewer/references/ for:

• evaluation_checklist.md - Complete skill evaluation criteria
• pr_template.md - Professional PR description template

---

github-contributor - GitHub Contribution Strategy

Strategic guide for becoming an effective GitHub contributor and building your open-source reputation.

When to use:

• Looking for projects to contribute to
• Learning contribution best practices
• Building your GitHub presence and reputation
• Understanding how to write high-quality PRs

Key features:

• Four contribution types: Documentation, Code Quality, Bug Fixes, Features
• Project selection criteria: What makes a good first project vs red flags
• PR excellence workflow: Before → During → After submission checklist
• Reputation building ladder: Documentation → Bug Fixes → Features → Maintainer
• GitHub CLI commands: Quick reference for fork, PR, issue operations
• Conventional commit format: Type, scope, description structure
• Common mistakes: What to avoid and best practices

Contribution types explained:

Level 1: Documentation fixes (lowest barrier, high impact)
    ↓ (build familiarity)
Level 2: Code quality (medium effort, demonstrates skill)
    ↓ (understand codebase)
Level 3: Bug fixes (high impact, builds trust)
    ↓ (trusted contributor)
Level 4: Feature additions (highest visibility)
    ↓ (potential maintainer)

Example usage:

# Install the skill
claude plugin install github-contributor@daymade-skills

# Find good first issues
"Help me find projects with good first issues in Python"

# Write a high-quality PR
"Guide me through creating a PR for this bug fix"

# Build contribution strategy
"Help me plan a contribution strategy for building my GitHub profile"

🎬 Live Demo

Coming soon

📚 Documentation: See github-contributor/references/ for:

• pr_checklist.md - Complete PR quality checklist
• project_evaluation.md - How to evaluate projects for contribution
• communication_templates.md - Issue and PR communication templates

---

i18n-expert - Internationalization & Localization

Complete internationalization/localization setup and auditing for UI codebases. Configure i18n frameworks, replace hard-coded strings with translation keys, ensure locale parity between en-US and zh-CN, and validate pluralization and formatting.

When to use:

• Setting up i18n for new React/Next.js/Vue applications
• Auditing existing i18n implementations for key parity and completeness
• Replacing hard-coded strings with translation keys
• Ensuring proper error code mapping to localized messages
• Validating pluralization, date/time/number formatting across locales
• Implementing language switching and SEO metadata localization

Key features:

• Library selection and setup (react-i18next, next-intl, vue-i18n)
• Key architecture and locale file organization (JSON, YAML, PO, XLIFF)
• Translation generation strategy (AI, professional, manual)
• Routing and language detection/switching
• SEO and metadata localization
• RTL support for applicable locales
• Key parity validation between en-US and zh-CN
• Pluralization and formatting validation
• Error code mapping to localized messages
• Bundled i18n_audit.py script for key usage extraction

Example usage:

# Install the skill
claude plugin install i18n-expert@daymade-skills

# Setup i18n for a new project
"Set up i18n for my React app with English and Chinese support"

# Audit existing i18n implementation
"Audit the i18n setup and find missing translation keys"

# Replace hard-coded strings
"Replace all hard-coded strings in this component with i18n keys"

🎬 Live Demo

Coming soon

📚 Documentation: See i18n-expert/SKILL.md for complete workflow and architecture guidance.

Requirements:

• Python 3.6+ (for audit script)
• React/Next.js/Vue (framework-specific i18n library)

---

claude-skills-troubleshooting - Plugin & Skill Troubleshooting

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-skills-troubleshooting)

Diagnose and resolve Claude Code plugin and skill configuration issues. Debug plugin installation, enablement, and activation problems with systematic workflows.

When to use:

• Plugins installed but not showing in available skills list
• Skills not activating as expected despite installation
• Troubleshooting enabledPlugins configuration in settings.json
• Debugging "plugin not working" or "skill not showing" issues
• Understanding plugin state architecture and lifecycle

Key features:

• Quick diagnosis via diagnostic script (detects installed vs enabled mismatch)
• Plugin state architecture documentation (installed_plugins.json vs settings.json)
• Marketplace cache freshness detection and update guidance
• Known GitHub issues tracking (#17832, #19696, #17089, #13543, #16260)
• Batch enable script for missing plugins from a marketplace
• Skills vs Commands architecture explanation
• Comprehensive diagnostic commands reference

Example usage:

# Run diagnostic
python3 scripts/diagnose_plugins.py

# Batch enable missing plugins
python3 scripts/enable_all_plugins.py daymade-skills

🎬 Live Demo

Coming soon

📚 Documentation: See claude-skills-troubleshooting/SKILL.md for complete troubleshooting workflow and architecture guidance.

Requirements: None (uses Claude Code built-in Python)

---

audio-router - StepFun Speech and Meeting-Minutes Routing

Install: claude plugin install daymade-audio@daymade-skills (suite-only — invoked as daymade-audio:audio-router)

Selects the installed StepFun ASR, StepFun TTS, or transcript-to-minutes specialist and reads its full instructions at use time. Their original slash commands remain available manually. General audio transcription and transcript correction retain direct automatic entries.

---

meeting-minutes-taker - Meeting Minutes Generator

Install: claude plugin install daymade-audio@daymade-skills (suite-only — invoked as daymade-audio:meeting-minutes-taker)

Transform meeting transcripts into high-fidelity, structured meeting minutes with iterative human review.

When to use:

• Meeting transcript provided and minutes/notes/summaries requested
• Multiple versions of meeting minutes need merging without content loss
• Existing minutes need review against original transcript for missing items

Key features:

• Multi-pass parallel generation with UNION merge strategy
• Evidence-based recording with speaker quotes
• Mermaid diagrams for architecture discussions
• Iterative human-in-the-loop refinement workflow
• Cross-AI comparison for bias reduction
• Completeness checklist for systematic review

Example usage:

# Install the full audio suite (includes meeting-minutes-taker)
claude plugin install daymade-audio@daymade-skills

# Then provide a meeting transcript and request minutes

🎬 Live Demo

Coming soon

📚 Documentation: See daymade-audio/meeting-minutes-taker/SKILL.md for complete workflow and template guidance.

Requirements: None

---

deep-research - Research Reports and Provider Runs

Generate research reports from a durable study record: original sources, claim-level citations, rejected leads, and prior studies remain available for the next question. When several AI products or modes investigate one decision, coordinate their existing Skills and agents in parallel, preserve original outputs, and synthesize against the underlying sources and business question.

When to use:

• Need a structured research report, literature review, or market/industry analysis
• Require strict section formatting or a template to be enforced
• Need evidence mapping, citations, and source quality review
• Want multi-pass synthesis to avoid missing key findings

Key features:

• Report spec and format contract workflow
• Evidence table with source quality rubric
• Multi-pass complete drafting with UNION merge
• Citation verification and conflict handling
• Project-local source and claim records for single-route and multi-route studies, with a searchable catalog for reuse
• Exact seed-document handoff, persistent session aliases, and verified-source-first catalog search
• Local provider × mode task and artifact ledger with provenance and hash checks
• Read-only parallel dispatch board that separates shared app control, active tasks, completed reports and held routes; actual provider calls follow their own Skills and authorization rules
• Ready-to-use report template and formatting rules

Example usage:

# Install the skill
claude plugin install deep-research@daymade-skills

# Then provide a report spec or template and request a deep research report

🎬 Live Demo

Coming soon

📚 Documentation: See deep-research/SKILL.md, research-asset-contract.md, and research_report_template.md.

Requirements: None

---

competitors-analysis - Evidence-Based Competitor Intelligence

Discover, clone, update, and analyze competitor repositories with evidence-based competitive intelligence. Repository-backed findings must come from local cloned code; market-landscape claims must cite their source and volatility.

When to use:

• Track and analyze competitor products or technologies
• Discover GitHub competitors for a product or market
• Create evidence-based competitor profiles
• Generate competitive landscape and opportunity reports
• Check whether competitor code has changed
• Need to document technical decisions with cited sources

Key features:

• Durable competitor workspace convention using $HOME/workspace/competitors/{product}/
• GitHub discovery workflow for shortlisting relevant repositories
• Repository ingest/update flow with remote + commit recording
• Required source citation format for repository facts (file:line_number)
• Landscape synthesis for positioning, strengths, weaknesses, opportunities, and risks
• Tech stack analysis guides for Node.js, Python, Rust projects
• Bundled templates: profile template, analysis checklist
• Management script for discover/clone-url/clone/pull/status operations

Example usage:

# Install the skill
claude plugin install competitors-analysis@daymade-skills

# Then ask Claude to analyze a competitor
"分析竞品 https://github.com/org/repo"
"添加竞品到 flowzero 产品的竞品列表"
"看看 claude-flow-viewer 这个方向最近有哪些竞品更新"

🎬 Live Demo

Coming soon

📚 Documentation: See competitors-analysis/SKILL.md and competitors-analysis/references/ for templates.

Requirements: Git (for cloning repositories)

---

tunnel-doctor - Tailscale + Proxy/VPN Conflict Fixer

Diagnose and fix conflicts when using Tailscale alongside proxy/VPN tools (Shadowrocket, Clash, Surge) on macOS. Covers four independent conflict layers with specific guidance for SSH access to WSL instances.

When to use:

• Tailscale ping works but SSH/TCP connections time out
• Proxy tools hijack the Tailscale CGNAT range (100.64.0.0/10)
• Browser returns HTTP 503 but curl and SSH work
• git push/pull fails with "failed to begin relaying via HTTP"
• Setting up Tailscale SSH to WSL and encountering operation not permitted
• Need to make Tailscale and Shadowrocket/Clash/Surge coexist on macOS

Key features:

• Four-layer diagnostic model: route hijacking, HTTP env vars, system proxy bypass, SSH ProxyCommand double tunneling
• Per-tool fix guides for Shadowrocket, Clash, and Surge
• SSH ProxyCommand double tunnel detection and fix (git push/pull failures)
• Tailscale SSH ACL configuration (check vs accept)
• WSL snap vs apt Tailscale installation (snap sandbox breaks SSH)
• Remote development SOP with proxy-safe Makefile patterns

Example usage:

# Install the skill
claude plugin install tunnel-doctor@daymade-skills

# Then ask Claude to diagnose
"Tailscale ping works but SSH times out"
"Fix Tailscale and Shadowrocket route conflict on macOS"
"git push fails with failed to begin relaying via HTTP"
"Set up Tailscale SSH to my WSL instance"

🎬 Live Demo

Coming soon

📚 Documentation: See tunnel-doctor/references/proxy_conflict_reference.md for per-tool configuration and conflict architecture.

---

windows-remote-desktop-connection-doctor - AVD/W365 Connection Quality Diagnostician

Diagnose Windows App (Microsoft Remote Desktop / Azure Virtual Desktop / W365) connection quality issues on macOS, with focus on transport protocol optimization (UDP Shortpath vs WebSocket fallback).

When to use:

• VDI connection is slow with high RTT (>100ms)
• Transport Protocol shows WebSocket instead of UDP
• RDP Shortpath fails to establish
• Connection quality degraded after changing network location
• Need to identify VPN/proxy interference with STUN/TURN

Key features:

• 5-step diagnostic workflow from connection info collection to fix verification
• Transport protocol analysis (UDP Shortpath > TCP > WebSocket hierarchy)
• VPN/proxy interference detection (ShadowRocket TUN mode, Tailscale exit node)
• Windows App log parsing for health check failures, certificate errors, FetchClientOptions timeouts
• ISP UDP restriction testing with STUN connectivity checks
• Chinese ISP-specific guidance for UDP throttling issues
• Working vs broken log comparison methodology

Example usage:

# Install the skill
claude plugin install windows-remote-desktop-connection-doctor@daymade-skills

# Then ask Claude to diagnose
"My VDI connection shows WebSocket instead of UDP, RTT is 165ms"
"Diagnose why RDP Shortpath is not working"
"Windows App transport protocol stuck on WebSocket"

🎬 Live Demo

Coming soon

📚 Documentation: See windows-remote-desktop-connection-doctor/references/ for log analysis patterns and AVD transport protocol details.

---

product-analysis - Multi-Path Product Analysis & Optimization

Audit UX, API, architecture and selected competitor implementations using parallel exploration and evidence-based synthesis.

When to use:

• Product launch readiness reviews
• Multi-perspective codebase and UX audits before release
• API quality checks with endpoint and consumption consistency reviews
• Competitive benchmarking against selected competitor repos

Key features:

• Auto-detects tool context (project stack + optional codex availability)
• Select the applicable audit scope
• Evidence verification and priority selection through the synthesis method
• Built-in comparison hooks with competitors-analysis

Example usage:

# Install the skill
claude plugin install product-analysis@daymade-skills

# Then ask Claude for analysis
"Run product-analysis in full mode for launch audit"
"Audit UX navigation and verify its effects on the user's task"
"Trace API consumers and identify routes safe to retire"
"Compare this product with our top competitors"

🎬 Live Demo

Coming soon

📚 Documentation: See product-analysis/SKILL.md and product-analysis/references/analysis_dimensions.md for dimension definitions and workflow guidance.

Requirements: Optional codex CLI (for multi-model parallel mode). Skill runs with Claude only if codex is not installed.

---

financial-data-collector - Financial Data Collection for US Equities

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:financial-data-collector)

Collect real financial data for any US publicly traded company from free public sources (yfinance). Output structured JSON with market data, historical financials (income statement, cash flow, balance sheet), WACC inputs, and analyst estimates - ready for downstream DCF modeling, comps analysis, or earnings review.

When to use:

• Collecting structured financial data before building DCF or valuation models
• Pulling market data (price, shares, beta, market cap) for any US equity ticker
• Gathering historical income statement, cash flow, and balance sheet data
• Getting risk-free rate (10Y Treasury) and analyst consensus estimates

Key features:

• Robust yfinance field mapping with alias chains (handles API instability across versions)
• NaN year detection and transparent reporting (never fills with estimates)
• 9-check validation: field completeness, market cap cross-check, CapEx sign convention, net debt consistency
• NO FALLBACK principle: missing data returns null with _source attribution, never default values
• FCF definition mismatch flagging (yfinance FCF ≠ investment bank FCF due to SBC)

Example usage:

# Install the suite
claude plugin install daymade-financial@daymade-skills

# Then ask Claude to collect data
"Collect financial data for META"
"Get financials for AAPL --years 3"
"Pull DCF inputs for NVDA"

🎬 Live Demo

Coming soon

📚 Documentation: See financial-data-collector/SKILL.md, output-schema.md, and yfinance-pitfalls.md.

Requirements: Python 3.11+, yfinance, pandas (auto-installed via uv inline dependencies).

---

excel-automation - Excel Creation, Parsing, and macOS Control

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:excel-automation)

Create professionally formatted Excel files, parse complex .xlsm models with stdlib XML/ZIP workflows, and control Microsoft Excel windows on macOS via AppleScript.

When to use:

• Building finance-ready spreadsheets with consistent formatting rules
• Parsing complex bank/broker .xlsm files that fail in openpyxl
• Extracting targeted sheet/cell data without loading huge workbooks
• Automating Excel window operations (zoom, scroll, select) on macOS

Key features:

• Production template for formatted workbook generation via openpyxl
• Complex workbook parser using zipfile + xml.etree (no heavy dependencies)
• Corrupted definedNames repair workflow for problematic files
• Verified AppleScript command patterns with timeout safeguards
• Bundled formatting reference for colors, number formats, and table patterns

Example usage:

# Install the documentation suite
claude plugin install daymade-docs@daymade-skills

# Then ask Claude to automate Excel workflows
"Create a formatted valuation template workbook"
"Parse this .xlsm and extract the DCF sheet"
"Generate an AppleScript sequence to zoom and scroll Excel before screenshot"

🎬 Live Demo

Coming soon

📚 Documentation: See excel-automation/SKILL.md and formatting-reference.md.

Requirements: Python 3.8+, uv, openpyxl (auto via uv run --with openpyxl), macOS for AppleScript window control.

---

capture-screen - Programmatic macOS Screenshot Capture

Install: claude plugin install daymade-macos@daymade-skills (suite-only — invoked as daymade-macos:capture-screen)

For desktop or app-window screenshots, start with the Skill's observation and control boundary, then select its capture route.

macos-permissions - Diagnose and Repair macOS Privacy Permissions

Install: claude plugin install daymade-macos@daymade-skills (suite-only — invoked as daymade-macos:macos-permissions)

For visible permission or Keychain dialogs, start with observation and classification. For protected-data failures without a visible dialog, enter the TCC request procedure.

---

continue-claude-code-work - Resume Interrupted Claude Work

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:continue-claude-code-work)

Continue a verified Claude Code Session without reopening the old interactive session. The continuation layer first consumes read-claude-code-history, then reconstructs the original business outcome, unfulfilled requests, corrections, proven prior assets, and one next action that directly advances the goal.

When to use:

• A user provides a Claude session ID and wants the task continued
• You need to inspect local .claude JSONL files instead of running claude --resume
• A previous session was interrupted and the next concrete step must be reconstructed
• A multi-agent workflow was interrupted and you need to know which subagents completed

Key features:

• Requires a chronological read receipt rather than separate user/assistant tail lists
• Restores original outcome, remaining work, rejected routes, and successful assets
• Verifies current files, git state, external writes, and background work before duplicating anything
• Makes the business result—not parser success, review completion, or subprocess exit—the completion unit

Example usage:

# Then ask Claude to resume from local artifacts
"continue work from session 123e4567-e89b-12d3-a456-426614174000"
"don't resume, just read the .claude files and continue"
"check what I was working on in the last session and keep going"

📚 Documentation: See continue-claude-code-work/SKILL.md.

Requirements: Python 3.8+, git for workspace reconciliation.

---

scrapling-skill - Reliable Scrapling CLI Workflows

Install, troubleshoot, and use Scrapling CLI with a verified static-first workflow for extracting HTML, Markdown, or text from webpages. Includes a diagnostic script for broken extras installs, Playwright browser runtime checks, and smoke tests against real URLs.

When to use:

• Users mention Scrapling, uv tool install scrapling, or scrapling extract
• You need to choose between static and browser-backed fetching
• You need to extract article bodies from WeChat public pages (mp.weixin.qq.com)
• A Scrapling install works partially but fails on missing extras, browser runtime, or TLS verification

Key features:

• Bundled diagnose_scrapling.py script for CLI, browser runtime, and live URL smoke tests
• Verified default path: start with extract get, escalate to extract fetch only when needed
• WeChat extraction pattern using #js_content for clean article Markdown
• Troubleshooting guidance for missing click, Playwright runtime setup, and curl: (60) trust-store failures
• Output validation workflow using file size and content checks instead of exit-code assumptions

Example usage:

# Install the skill
claude plugin install scrapling-skill@daymade-skills

# Then ask Claude to work through Scrapling for you
"Install Scrapling CLI and verify the setup"
"Extract this WeChat article into Markdown with Scrapling"
"Decide whether this page needs static or browser-backed fetching"

🎬 Live Demo

Coming soon

📚 Documentation: See scrapling-skill/SKILL.md and scrapling-skill/references/troubleshooting.md.

Requirements: Python 3.6+, uv, Scrapling CLI, and Playwright browser runtime for browser-backed fetches.

---

ima-copilot - Tencent IMA Companion & Installer

One-stop wrapper for the official Tencent IMA skill (ima.qq.com). Installs upstream ima-skill to Claude Code, Codex, and OpenClaw via npx skills add, guides API key setup, detects and repairs known upstream issues under user consent, and implements a personalized fan-out search strategy that floats priority knowledge bases to the top.

When to use:

• Users mention IMA, 腾讯 IMA, ima.qq.com, or need to install the official ima-skill
• Users report Skipped loading skill(s) due to invalid SKILL.md warnings related to ima-skill
• You need to search across IMA knowledge bases with KB-priority boosting
• You need to configure or rotate IMA API credentials
• Upstream ima-skill ships a known issue (e.g., missing YAML frontmatter in submodule files)

Key features:

• Zero-config installation to Claude Code / Codex / OpenClaw via vercel-labs/skills with auto-detection and default symlink mode (fix or upgrade once, every agent sees it)
• XDG-style credential management at ~/.config/ima/{client_id, api_key} with env-var fallback
• scripts/diagnose.sh read-only health check (install presence, credential liveness, known issues)
• scripts/search_fanout.py client-side cross-KB search with priority lists, subset-skip lists, and 100-hit silent-truncation detection
• Wrapper-only architecture: never vendors upstream files, never forks — every repair is a runtime instruction executed with explicit consent and automatic timestamped backups
• Two user-selectable repair strategies for the frontmatter issue (rename to MODULE.md or prepend minimal frontmatter)
• Personalization via ~/.config/ima/copilot.json with illustrative-only template values

Example usage:

# Install the skill
claude plugin install ima-copilot@daymade-skills

# Then ask Claude to drive the flow
"Install ima-skill and configure my IMA API key"
"Run diagnose on my ima-skill and fix whatever is broken"
"Search my IMA knowledge bases for embedding model comparisons, priority to my curated KB"

🎬 Live Demo

Coming soon

📚 Documentation: See ima-copilot/SKILL.md and ima-copilot/references/known_issues.md.

Requirements: Node.js 18+ (for npx skills), curl, unzip, Python 3.6+. IMA OpenAPI credentials from https://ima.qq.com/agent-interface.

---

claude-export-txt-better - Fix Claude Code Export Formatting

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:fixing-claude-export-conversations)

Reconstruct broken line wrapping in Claude Code exported .txt conversation files. Rebuilds tables, paragraphs, paths, and tool calls that were hard-wrapped at fixed column widths, and ships with an automated 53-check validation suite (file-agnostic, catches over- and under-merging regressions).

When to use:

• Users have a Claude Code export file where tables, paths, or tool output got mangled by line wrapping
• Users mention "fix export", "fix conversation", "make export readable"
• Users reference a file matching YYYY-MM-DD-HHMMSS-*.txt
• Users want to post-process /export output before sharing or archiving it

Key features:

• Deterministic Python script (fix-claude-export.py) with --stats mode for before/after metrics
• 53-check automated validator (validate-claude-export-fix.py) that catches regressions
• Evals directory with real fixture cases
• No external dependencies beyond uv and Python 3.8+

Example usage:

# Fix and show stats
uv run daymade-claude-code/claude-export-txt-better/scripts/fix-claude-export.py broken.txt --stats

# Custom output path
uv run daymade-claude-code/claude-export-txt-better/scripts/fix-claude-export.py broken.txt -o fixed.txt

# Validate the fix
uv run daymade-claude-code/claude-export-txt-better/scripts/validate-claude-export-fix.py broken.txt fixed.txt

🎬 Live Demo

Coming soon

📚 Documentation: See claude-export-txt-better/SKILL.md and the bundled evals/ fixtures.

Requirements: Python 3.8+, uv package manager.

---

douban-skill - Douban Collection Export & Sync

Export and sync Douban (豆瓣) book / movie / music / game collections to local CSV files via the reverse-engineered Frodo API. Full export covers all history; RSS incremental sync keeps daily updates current. No login, no cookies, no browser — just a user ID and it works.

When to use:

• Users want to back up their Douban reading/watching/listening/gaming history
• Users mention 豆瓣, douban, 读书记录, 观影记录, 书影音
• Users need incremental sync of recent Douban activity
• Users want CSV output compatible with Excel (UTF-8 BOM)

Key features:

• Full export of all 4 categories (books/movies/music/games) via Frodo API
• RSS incremental sync for daily updates (last ~10 items per feed)
• Pre-flight user-ID validation (fail-fast on wrong ID)
• UTF-8 BOM CSV output, Excel-compatible, cross-platform
• Bundled troubleshooting log documenting 7 tested scraping approaches and why each failed (Douban PoW challenges block every web-scraping approach — only Frodo API works)
• .gitleaks.toml allowlist for the public Android APK credentials

Example usage:

# Full export of user's collections
uv run douban-skill/scripts/douban-frodo-export.py <douban-user-id>

# Incremental RSS sync (last ~10 items per category)
uv run douban-skill/scripts/douban-rss-sync.py <douban-user-id>

🎬 Live Demo

Coming soon

📚 Documentation: See douban-skill/SKILL.md and douban-skill/references/troubleshooting.md for the complete failure log of rejected approaches.

Requirements: Python 3.8+, uv package manager. No login or cookies required.

---

terraform-skill - Terraform Operational Traps

Designs and diagnoses safe Terraform releases as well as the provisioner traps learned from real incidents. It keeps staging and production on one required configuration schema, validates exact candidate bytes + the Compose-rendered environment + the immutable runtime before live mutation, and binds saved plans to staging evidence, source provenance, explicit production authorization, and independent readback.

When to use:

• Writing null_resource provisioners or remote-exec blocks that SSH into fresh instances
• Setting up multi-environment (prod/staging/dev) Terraform with shared modules
• Debugging containers that are Restarting/unhealthy after terraform apply
• Hitting "docker: not found" in remote-exec, rsync connection drops in local-exec, or TLS cert errors
• Troubleshooting drift or provisioner failures during re-runs
• Configuring Caddy/gateway resources with Cloudflare credentials
• Reviewing a saved plan or broad deploy resource that may also mutate a shared gateway
• Closing staging/production config drift, receipt, provenance, or production-approval gaps

Key features:

• One required-key contract for every environment; values may differ, requiredness may not
• Exact-bundle prevalidation for every normal and recovery writer before any live write/restart
• Saved-plan, staging-receipt, remote-main provenance, production-approval, and live-readback gates
• Corrected provider/provisioner patterns for cloud-init, Docker, DNS, TLS, snapshots, and fresh hosts

Example usage:

# Trigger the skill naturally during Terraform work
"I'm getting 'docker: not found' in my null_resource provisioner after apply"
"My rsync local-exec is failing with 'connection unexpectedly closed'"
"Help me write a multi-env Terraform setup without snapshot cross-contamination"
"Staging has this Caddy variable but production leaves it empty — how do I validate both safely?"

🎬 Live Demo

Coming soon

📚 Documentation: See terraform-skill/SKILL.md and the bundled references/ for detailed remediation patterns.

Requirements: None (Terraform-adjacent knowledge only; no runtime dependencies).

---

slides-creator - Compatibility Notice

Read the retirement and installation-compatibility notice before using the legacy slides-creator entry.

---

excalidraw-use - Place Images onto an Excalidraw Board

Batch-place existing images onto an Excalidraw whiteboard, laid out on a generous grid so nobody has to drag them apart afterwards. Also turns a slide deck into clean per-slide images first, and inspects what is inside a .excalidraw file. The available Excalidraw MCP servers and skills cover element CRUD and export but document no image element type, no dataURL handling, and no files map — embedding your own pictures is the gap this fills.

When to use:

• Putting screenshots, a picture library, or deck slides onto a whiteboard
• Spacing many images out so they never need manual adjustment
• Turning a Vite/React slide deck into images you can draw over
• Inspecting a scene file: element mix, embedded payload size, occupied extent

Key features:

• Content-hash dedupe and --exclude for images already on the board
• --template-from copies the image-element field set out of your own board — Excalidraw's published schema stops before fileId/status/scale/crop
• Write-back verification: fails on a missing file entry, a distorted aspect ratio, or any overlap
• Deck capture hides presenter chrome, expands staged reveals, and reports fragments that never rendered
• Documents the two silent destroyers: Open and drag-and-drop replace a scene (only the clipboard merges), and a stale build removes a source feature while innerText still reads hidden fragments as present

Example usage:

# Trigger the skill naturally
"Put these screenshots on my Excalidraw board"
"Add my old workshop images to the canvas, spaced out"
"Turn this deck into images I can draw on"

Note: Not for generating a diagram from a text description — that is a different job.

---

debugging-network-issues - Evidence-Driven Network Investigation

Falsification-first methodology for network, streaming, and protocol-layer bugs where the obvious cause is probably wrong. Built from a real 5-hour SSE incident where assumption-stacking wasted hours that a 10-minute layered experiment would have resolved.

When to use:

• Connection resets (ECONNRESET, HTTP/2 RST_STREAM, INTERNAL_ERROR)
• SSE / long-polling stalls or fixed-time drops (60s, 100s, 130s)
• CDN / proxy / CGNAT idle-timeout incidents
• Client-side proxy / VPN / TUN misrouting (e.g. ERR_CONNECTION_CLOSED, SSL_ERROR_SYSCALL, fake TUN DNS IPs, CNAME-based rule overrides)
• Certificate-verification errors (UNKNOWN_CERTIFICATE_VERIFICATION_ERROR, wrong-site certificate)
• Any "works sometimes / fails after N seconds" pattern
• LAN-layer mysteries: unknown devices on the local network, devices silenced by a subnet change, hosts "dead" on one segment but alive on another
• Multi-hop systems (client → CDN → LB → reverse proxy → app → upstream) where a symptom could plausibly come from several layers

Key features:

• Layered isolation experiments: run the same logical request through three or more paths differing by exactly one hop
• Env-gated runtime instrumentation patterns (no production-code mutation)
• Counter-review four-question filter to challenge single-cause assumptions
• Bundled probe scripts (layered-isolation-probe.sh, mock-idle-upstream.py)
• Real case studies: SSE RST_STREAM at 130s caused by CGNAT idle timeout; proxy/TUN CNAME rule override causing ERR_CONNECTION_CLOSED

Requirements: None (methodology + portable shell/Python probes).

---

stepfun-tts - StepFun StepAudio 2.5 Contextual TTS

Install: claude plugin install daymade-audio@daymade-skills (suite-only — invoked as daymade-audio:stepfun-tts)

Generate Chinese / Japanese speech with stepaudio-2.5-tts. Captures the two non-obvious TTS pitfalls that cost hours otherwise: voice_label removal (replaced by natural-language instruction) and stricter 2.5-era censorship (死/消失/political terms).

When to use:

• Chinese / Japanese TTS with emotional and prosody control (whisper, pause, stress, mid-sentence pivot)
• Batch-generating game / app voice lines with per-line censorship_block fallback
• Migration from step-tts-2 to stepaudio-2.5-tts (voice_label → instruction breaking change)
• Hitting StepFun censorship blocks on previously-fine content

Key features:

• stepaudio-2.5-tts with instruction (≤200 chars natural-language mood) + inline () prosody
• Bundled tts_generate.py (with --batch <jsonl>) and ab_compare.sh
• API key resolution: $STEPFUN_API_KEY → ${CLAUDE_PLUGIN_DATA}/config.json fallback
• Censorship rewrite playbook in references/migration_from_v2.md

Requirements: StepFun API key, "Normal" tier (https://platform.stepfun.com/). For ASR / transcription, use the sibling stepfun-asr skill below.

---

stepfun-asr - StepFun StepAudio 2.5 ASR (SSE Endpoint)

Install: claude plugin install daymade-audio@daymade-skills (suite-only — invoked as daymade-audio:stepfun-asr)

Transcribe Chinese / English audio with stepaudio-2.5-asr. Hides the #1 trap of the 2.5 ASR family: it does NOT live on /v1/audio/transcriptions — the wrong endpoint returns a misleading model stepaudio-2.5-asr not supported error that looks identical to a permission/whitelist failure.

When to use:

• Long audio transcription (up to ~30 minutes single-call, 32K context, ~85-101× RTF — no client-side chunking)
• Migration from step-asr / step-asr-1.1 (different endpoint, different body shape, SSE response)
• Hitting the misleading model stepaudio-2.5-asr not supported error (= wrong endpoint, not permission)
• Silent 4xx auth failures on audio endpoints (= using a "Plan" key instead of a "Normal" key)

Key features:

• /v1/audio/asr/sse SSE streaming with base64 audio + nested JSON body (the script handles all four traps)
• Bundled asr_transcribe.py — pure-stdlib CLI, auto-detects mp3/wav/ogg/opus/pcm by extension
• Handles SSE error events (censorship can fire on ASR side too — rare but real)
• API key resolution: $STEPFUN_API_KEY → ${CLAUDE_PLUGIN_DATA}/config.json fallback
• Suggests transcript-fixer (ASR error correction) and meeting-minutes-taker (structured minutes) as natural downstream skills

Requirements: StepFun API key, "Normal" tier (https://platform.stepfun.com/). Plan keys cannot call audio endpoints.

---

auto-repo-setup - Automated Repository Setup & Environment Repair

Make a repository runnable and handoff-ready without guessing its stack or changing how collaborators normally work. The skill reads project authority first, repairs the verified gap, and treats startup instructions, lifecycle hooks, and Git mutation as different mechanisms with different safety boundaries.

When to use:

• Someone says "跑不起来", "怎么启动", "环境怎么配", or "帮我设置代码库"
• Setting up a new machine or creating a durable repository handoff
• Adding routine startup sync for Claude Code/Codex without auto-stashing local work
• Diagnosing repeated SessionStart output before changing hook configuration
• Adding a lifecycle hook only when behavior must occur before the first prompt
• Sanitizing git history after accidental secret/path leaks
• Handling merge conflicts or git push failures with explicit safety gates

Key features:

• Outcome router: separates environment repair, routine sync, handoff, hook diagnosis, and explicit pre-prompt automation
• Stack-aware inventory: check_env.py infers only declared toolchains from manifests/lockfiles; it does not assume ffmpeg, uv, Python, or .env
• Startup boundary: defaults stable behavior to AGENTS.md/CLAUDE.md or a normal Agent request; the Claude hook manager is guarded, dry-runnable, idempotent, and removable
• Safety guardrails: Push Safety (visibility verification before any push), PII Guard (4-layer secret scanning), NO FALLBACK principle for env vars, Git Hook Bypass ban
• Counter-review boundary: reserves multi-agent review for material shared-config/security/destructive changes, not ordinary setup checks
• Bundled scripts: stack-aware inventory, guarded Claude startup-nudge manager, and read-only history candidate scan

Example usage:

# Install the skill
claude plugin install auto-repo-setup@daymade-skills

# Then ask Claude naturally
"我跑不起来这个仓库"
"帮我设置一下这个项目的环境"
"进入项目先同步远端；本地有改动不要自动 stash"
"为什么同一个 SessionStart 输出了三次？先查清来源"
"只有首条消息前必须注入动态提醒时，才帮我装 hook"
"git push 被拒了"

Requirements: The guidance itself has no runtime dependency. Bundled Python utilities require Python 3.10+; no external API key is required.

---

terminal-screenshot - See the Real Visual Result of Terminal Output

Render a terminal CLI program's colored output to a PNG so Claude can actually see the rendered result — color contrast, alignment, background blocks, highlighting — instead of only reading plain text and raw ANSI escape codes. Reading a hex value is guessing; seeing the rendered contrast on the real terminal background is verification.

When to use:

• Right after changing any CLI color config (delta / bat / themes / lazygit pager) to visually confirm the result
• Verifying git diff (delta) add/remove contrast, bat syntax highlighting, starship prompt, eza/ls colors, ripgrep matches
• Any time you need to judge "does this color look right / is the contrast enough" instead of guessing from hex codes

Key features:

• Capture-then-render discipline: captures full-fidelity ANSI in a normal shell first, then renders — never lets the renderer run complex CLIs (which degrade in a child pty and drop background blocks)
• freeze-first, zero-dependency fallback: prefers charmbracelet/freeze for faithful rendering; falls back to a bundled stdlib ANSI→HTML converter + headless Chrome when freeze is unavailable
• Real terminal background: renders on the actual terminal background color so dark themes are judged accurately
• Per-CLI capture recipes: delta, git, bat, eza, ls, ripgrep, and a generic forced-color path
• Bundled scripts: render_ansi.sh (freeze/Chrome auto-select), ansi2html.py (stdlib renderer)

Example usage:

# terminal-screenshot lives in the daymade-claude-code suite
claude plugin install daymade-claude-code@daymade-skills

# Then ask Claude naturally
"verify my delta diff colors"
"看一下这个终端配色的真实效果"
"is the add/remove contrast in git diff strong enough?"

Requirements: macOS. charmbracelet/freeze (preferred renderer) or Google Chrome (fallback). Python 3 for the fallback renderer.

---

pdf-to-html - Read a PDF as Faithful HTML (with Optional Translation)

Convert a PDF into one self-contained, readable HTML file that preserves images, charts and reading order — optionally translating it into another language while keeping every figure. A PDF is a layout, not just a text stream, so the workflow renders each page for you to see before building, and renders the HTML for visual verification before delivery.

When to use:

• Reading a PDF as a clean web page or document (especially on a phone)
• Turning a report or whitepaper PDF into styled HTML without losing its figures
• Translating a PDF into another language while keeping its images, charts and tables in place

Key features:

• Structured extraction (PyMuPDF): text blocks with font sizes + images, with decorative images (footer logos, rules) auto-detected and dropped
• Data-driven build: heading levels inferred from font size, content images compressed and base64-inlined into one portable file
• Optional parallel translation: a Dynamic Workflow translates pages concurrently, captions data charts, and reconciles terminology — with fidelity rules (never invent a translated name; copy numbers and proper nouns verbatim)
• Mandatory visual verification: adaptive headless-Chrome screenshot sliced into readable segments (works around Chrome's ~16384px screenshot cap)
• Bundled failure-cases reference: the real traps (verification, rendering limits, fidelity) so they are not re-discovered

Example usage:

# pdf-to-html lives in the daymade-docs suite
claude plugin install daymade-docs@daymade-skills

# Then ask Claude naturally
"把这个 PDF 转成中文网页版"
"make this report readable as HTML"
"translate this PDF to English but keep the charts"

Requirements: uv, Google Chrome or Chromium (visual verification). Python packages (PyMuPDF, Pillow, numpy) auto-install via uv run --with.

---

asr-transcribe-to-text - Audio/Video Transcription with Qwen3-ASR

Install: claude plugin install daymade-audio@daymade-skills (suite-only — invoked as daymade-audio:asr-transcribe-to-text)

Transcribe audio and video files to text using Qwen3-ASR via two interchangeable inference paths: local MLX on macOS Apple Silicon (no API key, 15-27x realtime) or a remote vLLM/OpenAI-compatible API for any platform. Auto-detects the platform and recommends the best path, persisting the choice in ${CLAUDE_PLUGIN_DATA}/config.json.

When to use:

• Transcribing meeting recordings, lectures, interviews, podcasts, or screen recordings
• Converting any audio/video file to text (speech-to-text)
• Local, free transcription on an Apple Silicon Mac, or remote API when local is unavailable
• The first stage of a transcribe → correct → minutes pipeline

Key features:

• Dual inference paths — local MLX (15-27x realtime, free) and remote API, with automatic platform detection
• Bundled transcribe_local_mlx.py loads the model once and processes files sequentially (no GPU contention)
• Uses low-energy ~20-minute chunks with an 8192-token per-chunk bound, atomic checkpoints/resume, and full process-tree cleanup so one bad chunk cannot become an unbounded GPU job
• Remote fallback overlap_merge_transcribe.py splits into 18-minute chunks with 2-minute overlap and fuzzy-merges
• ffmpeg video→16kHz mono WAV extraction, truncation verification, and proxy-bypass handling
• Proactively suggests transcript-fixer to clean ASR recognition errors on the output

Example usage:

# asr-transcribe-to-text lives in the daymade-audio suite
claude plugin install daymade-audio@daymade-skills

# Then ask Claude naturally
"transcribe this meeting recording to text"
"把这个录音转成文字"
"convert lecture.mp4 to a transcript"

Requirements: uv, ffmpeg/ffprobe. Local MLX path needs macOS Apple Silicon; remote path needs a reachable vLLM/OpenAI-compatible ASR endpoint. No API key for local mode.

---

marketplace-dev - Skills Repo → Plugin Marketplace

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:marketplace-dev)

Create and maintain Claude Code plugin marketplaces: convert a repository, consolidate standalone skills into a new or existing suite, move skills between suites, validate real installation/cache boundaries, and ship the result through a PR.

When to use:

• Making a skills repo installable via claude plugin install
• Generating or fixing a marketplace.json (plugin distribution, one-click install, auto-update)
• Adding a new plugin to an existing marketplace and bumping the right versions
• Putting existing skills into a suite, moving a skill between suites, or making members suite-only
• Debugging schema rejections like Unrecognized key: "$schema" or duplicate plugin names

Key features:

• Evidence-intake phase that mines docs and local session history instead of guessing from a template
• Encodes non-obvious schema rules: $schema is rejected, metadata has only 3 valid fields, strict: false semantics, single-skill vs suite source/skills patterns
• Bundled check_marketplace.sh runs four checks (JSON syntax → claude plugin validate → source/skills resolution → reverse sync) and exits non-zero on failure
• Installation, cache-footprint, and GitHub-install test recipes to confirm source produced the intended snapshot
• Dedicated suite-consolidation workflow covering canonical moves, byte/mode preservation, repository-wide install/path drift, existing-user migration, isolated real installation, and immutable review
• Two PostToolUse hooks (validate on marketplace.json edit; warn on un-bumped version when a SKILL.md changes) that auto-activate with the plugin

Example usage:

# marketplace-dev lives in the daymade-claude-code suite
claude plugin install daymade-claude-code@daymade-skills

# Then ask Claude naturally
"turn this skills repo into a plugin marketplace"
"generate a marketplace.json for this repo and validate it"
"add my new skill to the marketplace and open a PR"
"move these standalone skills into daymade-macos and make them suite-only"

Requirements: claude CLI (for claude plugin validate / install tests), jq. Git remotes configured if opening an upstream PR.

---

skill-creator - Create, Improve & Benchmark Skills

Install: claude plugin install daymade-skill@daymade-skills (suite-only — invoked as daymade-skill:skill-creator)

The essential meta-skill for building your own skills. It scales verification to the change: bounded fixes get targeted checks, narrow behavior changes get sampled replays, and broad/high-risk work becomes eligible for heavier evidence without automatically starting it. Paired baselines, graders, benchmarks, and viewers require explicit authorization or a decision-bearing evidence plan plus opt-in. It also supports explicit benchmarking and optimizes a skill's description for better triggering accuracy.

When to use:

• Creating a skill from scratch, or editing/optimizing an existing one
• Running evals to test a skill, or benchmarking performance with variance analysis
• Improving a skill's description so Claude triggers it more reliably
• Wrapping a third-party CLI tool you just got working into a reusable companion skill

Key features:

• Prior-art research across the live conversation, explicitly approved prior history, local SOPs, installed plugins/MCPs, skills.sh, official plugins, npm/PyPI — to reuse infrastructure and encode only the user's unique methodology
• The inline-vs-context: fork decision guide (subagents can't spawn subagents or call skills) and composable/orthogonal skill design
• init_skill.py scaffolding, package_skill.py (auto-validates), and security_scan.py (gitleaks-based secret/PII detection)
• Existing-skill migration gate: tool-attested snapshot or verified Git-commit baseline, runtime-reachability-aware capability audit, explicit dispositions, and package-time re-verification that a clean commit or hand-written marker cannot bypass
• Risk-scaled verification router: Tier 1 targeted checks, Tier 2 sampled behavior replay, and Tier 3 broad/high-risk classification without automatic fan-out
• Separately authorized full-eval harness: with-skill + baseline runs, assertions, grading, benchmark aggregation, and an HTML viewer after an explicit request, or after a decision-bearing evidence plan receives opt-in
• Mandatory sanitization read-through for public skills — catches no-keyword leaks scanners miss
• Description-optimization loop (60/40 train/test split, selects best description by held-out score)

Example usage:

# skill-creator lives in the daymade-skill suite
claude plugin install daymade-skill@daymade-skills

# Then ask Claude naturally
"create a skill that does X"
"improve this skill's description so it triggers more reliably"
"benchmark this skill against a no-skill baseline"

Requirements: Python 3, uv, PyYAML (validation/packaging), gitleaks (security scan). claude CLI only for agent evals and description-optimization runs.

---

feishu-doc-scraper - Feishu/Lark → Faithful Markdown + Source-First Archive

For supported inputs, extraction steps and prerequisites, read the Skill guide. For archiving, follow the destination handoff and storage contract.

Example usage:

# Install the skill
claude plugin install feishu-doc-scraper@daymade-skills

# Then ask Claude naturally
"把这个飞书合集导出成 markdown"
"export this Feishu Minutes transcript"
"save this Lark wiki page as Markdown"

---

financial-router - Financial Data Routing

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:financial-router)

Routes Bigdata/RavenPack, free US company fundamentals, A-share news, sector Top N and announcements, and Gangtise setup to one bundled specialist. Pharmaceutical reporting and general argument review keep their direct Skills; all specialist slash commands remain available for manual use.

---

bigdata-skill - Bigdata.com (RavenPack) SDK + REST Toolkit

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:bigdata-skill)

Pull Bigdata.com (RavenPack) financial and news data through the official bigdata-client SDK and its public /v1/* REST endpoints — reaching the structured substrate the Bigdata MCP server doesn't hand over. The MCP returns prose chunks and pre-synthesized tearsheets; this toolkit reaches structured financials, prices, analyst estimates, a daily entity-sentiment series, annotated chunk search with sentiment + entity spans, and a screener.

When to use:

• Using Bigdata.com / RavenPack and the MCP result feels thin ("where's the sentiment score?", "I need entity-level data", "the calendar")
• Pulling forward/structured financials: analyst estimates, earnings/event calendar, surprises, ratings, price targets, statements, TTM metrics, a company screener
• Wanting annotated news chunks with numeric sentiment + entity spans, a sentiment time series, or a co-mention graph
• Mentions a bd_v2_ API key, rp_entity_id, query_unit/chunk cost, bigdata-client, or "the bigdata MCP isn't enough"

Key features:

• One BigdataClient exposing both the SDK (search + knowledge graph) and a REST escape hatch (bd._api.http) for every /v1/* endpoint the SDK never wrapped
• Routing table mapping each question to the right module; fields_values_to_records() to flatten {fields, values} responses
• Cost discipline: 1 query_unit = 10 chunks, only chunk-search billed, ChunkLimit (never a bare int), rerank thresholds, 50%-cheaper batch search, and a CostModel/CostTracker budget veto
• The "two data faces" guidance — structured financial (works for A-shares via English name/ISIN) vs unstructured Chinese NLP (a data-source-level dead end)
• rc() SSL-retry wrapper for the common first-handshake SSL: UNEXPECTED_EOF, plus a known-pitfalls reference with reproductions and fixes
• Fail-fast on a missing BIGDATA_API_KEY (no plaintext fallback); read-only, never writes/uploads

Example usage:

# Install the suite
claude plugin install daymade-financial@daymade-skills
export BIGDATA_API_KEY=bd_v2_xxxxxxxx

# Then ask Claude naturally
"pull NVIDIA's forward analyst estimates and last earnings surprise from Bigdata"
"give me a daily entity-sentiment series for this ticker"
"the bigdata MCP only gave me a tearsheet — I need the structured fields"

Requirements: A bd_v2_ Bigdata.com API key (env var, never hardcoded), uv, the official bigdata-client SDK in an isolated venv. Optional outbound/WSS proxy only if your network needs one to reach api.bigdata.com.

---

gangtise-copilot - Gangtise Investment-Research Suite Installer

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:gangtise-copilot)

One-command installer, credential configurator, and diagnostic layer for the full Gangtise (岗底斯投研) OpenAPI skill suite. Installs the official Gangtise skills (data, research, utility), configures accessKey/secretAccessKey with a live auth check, and runs a read-only health diagnostic — solving the suite's core discoverability problem (no public manifest, listing-disabled OBS bucket, parallel naming lines).

When to use:

• The user mentions Gangtise / 岗底斯, or any gangtise-* skill
• Setting up Gangtise credentials (accessKey / secretAccessKey)
• Errors like token is invalid / 接口地址错误, or "my gangtise install is broken"
• Routing a data question (research reports, chief-analyst opinions, OHLC, valuation) to the right Gangtise skill

Key features:

• install_gangtise.sh downloads the declared OBS bundles → extracts skill directories → symlinks them into detected agent skills dirs (Claude Code, OpenClaw, Codex), with minimal/workshop/full/--only presets
• configure_auth.sh writes a shared XDG credential file (mode 600), runs a live auth call, and symlinks every skill's .authorization to it
• Read-only diagnose.sh reports install state, credential validity, and scoped capability tiers (auth scope vs RAG scope)
• Skill registry routing a data question by data tier and operation type
• Wrapper contract: never vendors/forks upstream files, always re-downloads the canonical OBS artifact, and asks before touching any installed skill

Example usage:

# Install the suite
claude plugin install daymade-financial@daymade-skills

# Then ask Claude naturally
"装一下 gangtise 的所有 skill 并配置好凭据"
"my gangtise skills report token is invalid — diagnose it"
"宁德时代的研报用哪个 gangtise skill 查"

Requirements: A Gangtise accessKey + secretAccessKey; bash, curl, network access to the official OBS bucket and open.gangtise.com. Works with Claude Code, OpenClaw, and Codex agent layouts.

---

llm-wiki-setup - Co-Create a Personal Investment-Research LLM Wiki

Co-create a personal investment-research LLM Wiki (Andrej Karpathy's pattern) where the user's OWN analysis framework becomes a living CLAUDE.md — built by interviewing them rather than handing over a template. Pure markdown + [[wikilinks]]; knowledge is compiled into pages instead of being re-retrieved from raw files on every query (Karpathy's core idea — no RAG pipeline in place of compilation; a local search engine such as qmd over the grown wiki is part of his pattern). The value is extracting the user's personal investment preferences into THEIR OWN schema, never imposing a standard one.

When to use:

• Building a compounding research knowledge base (投研第二大脑 / 投研知识库 / 个人投研 wiki)
• Instantiating Karpathy's LLM Wiki pattern for finance/investing
• Turning a stock-picking, analyst-tracking, or earnings-watching workflow into a structured markdown vault
• Ingesting research reports / earnings calls / expert notes into an existing wiki, or running post-earnings prediction→fulfillment reviews

Key features:

• Sharp mechanism-layer vs rule-layer split: the three-level directory + wikilink + lint + git hook scaffold is copyable; the analysis schema is interview-grown, never templated
• init_vault.py scaffolds the mechanism layer only (no schema), then an 8-dimension interview builds the user's own CLAUDE.md in their own words
• Anti-corrosion: git hook + lint-vault.py keep the vault consistent and fight derived-value drift
• SOPs for ingesting a real source (HITL 5-checkpoint flow) and post-earnings fulfillment reviews
• Runs inline (calls the analyst-track-record skill and Bash) and chains into analyst-track-record for analyst back-testing — without rebuilding it

Example usage:

# Install the skill
claude plugin install llm-wiki-setup@daymade-skills

# Then ask Claude naturally
"帮我搭一个投研第二大脑"
"build me a personal investment-research wiki in Karpathy's style"
"ingest this earnings call into my research vault"

Requirements: Python 3, uv (for init_vault.py / lint), git. Markdown + wikilinks; no vector DB or embedding service required (qmd is an optional local search layer once the wiki grows). Pairs with the analyst-track-record skill for back-testing.

---

benchmark-due-diligence - Adversarial Teardown of an Envied Benchmark

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:benchmark-due-diligence)

Run adversarial due-diligence on a benchmark the user envies — a founder, KOL, company, or product whose claimed success looks inflated — separating marketing bubble from real signal, then mapping the validated playbook onto the user's own resources. The adversarial, decision-oriented cousin of deep-research: it assumes the picture is inflated until proven otherwise and ends in "what this means for ME", not a neutral report.

When to use:

• Wanting to 尽调/对标/拆解 a competitor or role-model, or 抄/偷师 someone's playbook
• Suspecting 水分/泡沫 in someone's claims (#1 on Product Hunt, 0-to-1M users, funding, 估值几个亿)
• Asking whether wins are 真本事 vs 运气/时机, or saying someone is 太成功了 and wanting the real story
• Preferring a debunk + replicable playbook over deep-research's neutral briefing

Key features:

• Two strictly-separated injection channels — public FACTS go to every agent; private COMMISSIONER_CONTEXT reaches only the final mapping agent (so client names never leak into open-web searches)
• Phase 0 foundation-by-evidence: verifies the benchmark's real entity graph and headline-claim attribution before any fan-out (don't reason from names/domains)
• Four-phase orchestration — collect → adversarial verify (L1-L4 grading, 坐实/存疑/证伪-水分 verdicts) → due-diligence conclusion (bubble-busting table + attribution breakdown) → commissioner resource-mapping
• Reuses existing plumbing instead of rebuilding it (deep-research fan-out, osint-investigate identity checks, the qcc family for 工商 data, agent-reach for social-platform data)
• Runs inline (it's an orchestrator — context: fork would silently break the fan-out)

Example usage:

# Install the financial research suite
claude plugin install daymade-financial@daymade-skills

# Then ask Claude naturally
"帮我尽调一下这个创始人，他到底有没有水分"
"tear down this competitor's playbook and tell me what I can actually copy"
"this KOL claims 0-to-1M users — is that real, and is it replicable for me?"

Requirements: Web access for the collection/verification agents. Optionally composes with deep-research, osint-investigate, the qcc skill family, and agent-reach; renders a shareable report via pdf-creator.

---

bilibili-source - Login-Free Bilibili Video Data + Danmaku Fetcher

Fetch real, citable data for any Bilibili (B站) video — title, UP follower count, publish date, tags, partition, per-part cids, live stats (view/like/coin/favorite/share/reply/danmaku), and full danmaku (bullet-comment) text — in one view/detail call, login-free. Built so engagement numbers are cheap to fetch and impossible to fake, instead of hand-typed into a doc where they rot.

When to use:

• Ingesting a Bilibili video into a knowledge base, or building a "why did this perform" case study
• Verifying a creator's claimed view/like/favorite numbers, or about to write any B站 metric into a document
• Wanting the danmaku text (qualitative audience reactions), not just a reply count
• Pasting a BVID, av number, b23.tv short link, or full URL — all normalized automatically

Key features:

• One bili-fetch.sh returns full metadata + live stats + UP fans + tags + every part's cid; metrics carry a fetched_at timestamp because they drift in real time
• bili-danmaku.sh pulls and decompresses the danmaku full text; bili-subs.sh handles the login-gated subtitle track (asks before touching browser cookies)
• bili-selftest.sh health-check verifies every endpoint against the live API, so API drift surfaces as one clear FAIL instead of a silent wrong answer
• NO-FABRICATION discipline: an unfetchable number is marked unverified, never estimated
• Strips the local proxy (Bilibili is a domestic CN service), sends UA+Referer (avoids HTTP 412), retries with backoff
• API reference includes the login-gated favorites (收藏夹) enumeration endpoints (x/v3/fav/*, verified no-WBI), the verified SESSDATA subtitle path, and the WBI request-signing algorithm for space/wbi/* extension

Example usage:

# Install the skill
claude plugin install bilibili-source@daymade-skills

# Then ask Claude naturally
"pull the real view/like/favorite counts for this B站 video so I can cite them"
"这个 B站 视频弹幕里大家在说什么？"
"grab the subtitle transcript from this bilibili video so I can summarize it"

Requirements: curl, jq, python3 (danmaku decompression). yt-dlp only for the login-gated subtitle path. No login for stats / metadata / danmaku.

---

ghostty-use - Ghostty Terminal Session Snapshot & Restore

Preserve and recover recorded Ghostty CLI sessions through the canonical Skill. For scheduled metadata backups, partial coverage and operating prerequisites, use the automatic workflow.

claude plugin install ghostty-use@daymade-skills

# Then ask Claude naturally
"我要重启电脑更新，把终端会话都保存一下"
"恢复之前的 ghostty 窗口和会话"
"restore my ghostty tabs after the update"
"检查一下有没有会话丢了"

---

claude-usage-analyst - Explain Claude Code Token Usage & Quota Burn

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-usage-analyst)

Turn local ccusage data into an evidence-based, human-readable explanation of where your Claude Code / Claude Desktop tokens, cost, and quota went — separating observed numbers from interpretation instead of guessing.

When to use:

• Asking why a Claude quota or 5-hour block got exhausted
• Wondering whether a model (fable / opus / sonnet) is unusually expensive for your workload
• Needing today's or a historical window's token/cost breakdown, including cache read/write pressure
• Explaining usage to a non-technical reader without unexplained jargon

Key features:

• Bundled analyze_claude_usage.py summarizes tokens, cost, input/output, and cache create/read over any date window and timezone
• Model-comparison mode (--model-a / --model-b) weighs both token volume and estimated cost — a model can be cheap per token but expensive overall
• A 5-hour-block table for quota-exhaustion questions
• Evidence discipline: every numeric claim is grounded in ccusage output; cache-read pressure is counted even when you never typed those tokens
• Scope is stated explicitly: ccusage claude measures local Claude Code logs, not a full Claude.ai chat bill

Example usage:

# Install the suite
claude plugin install daymade-claude-code@daymade-skills

# Then ask Claude naturally
"why did my Claude quota run out today?"
"is opus more expensive than sonnet for what I'm doing?"
"break down my Claude Code token usage for this month"

Requirements: ccusage (via npm i -g ccusage or npx ccusage@latest), python3.

---

marketplace-health-check - Full 6-Dimension Repo Health Check

claude plugin install marketplace-health-check@daymade-skills

Run a comprehensive, evidence-based health check of this skills marketplace repo with a parallel fan-out Dynamic Workflow — six inspectors cover code/script safety, documentation/SSOT consistency, security/PII leaks, open-PR triage, open-issue triage, and marketplace-manifest integrity at once — then the serious findings are Counter-Reviewed before they reach the report.

When to use:

• Before a release, or any time you want a full "is this whole repo OK across the board" sweep
• Checking whether docs/versions are consistent, PRs/issues are triaged, or PII has leaked into a public skill
• 全面体检 / 检查仓库状态 / 审计一下仓库

Key features:

• Six parallel inspectors (one per dimension) via a Dynamic Workflow — fast and focused (~15-20 min)
• Counter-Review: every high/critical finding is verified by hand before it reaches the report (agent findings are hypotheses, not conclusions) — catches false alarms AND wrong fixes
• Priority-ranked report: must-fix / backlog / optional / key insights, each item tagged real vs false-alarm
• Bundles the proven workflow script + a methodology reference (anti-target PII rule, working-copy-vs-history, scan-marker necessary-not-sufficient, the broken-install-command bug class)
• Inline orchestrator — drives the Workflow tool, so it never runs forked

Example usage:

# Install
claude plugin install marketplace-health-check@daymade-skills

# Then ask Claude naturally
"do a full health check of this repo before I cut a release"
"audit the marketplace — code, docs, PII, PRs, issues, everything"
"全面体检一下这个仓库"

Requirements: gh CLI (authenticated), git, jq, python3; opt-in to the Workflow tool (asking to run the health check is the opt-in).

---

claude-switch-models-setup - Multi-Provider Claude Code Profiles

claude plugin install daymade-claude-code@daymade-skills

Set up multiple isolated Claude Code CLI profiles so you can run different LLM providers (Kimi, MiniMax, GLM, DeepSeek, StepFun, Anthropic) in separate terminal windows at the same time — each profile gets its own .claude.json state while sharing skills, projects, hooks, and agents.

When to use:

• You want one terminal with Kimi and another with DeepSeek running side-by-side
• You need to switch between Anthropic and third-party models without config bleed
• You're setting up a post-workshop environment for students who want the same multi-provider workflow

Key features:

• One-click installer links its declared runtime scripts into ~/.config/claude-switch-models-setup/ and seeds an empty Codex activation manifest only when none exists
• Includes provider-specific ~/.claude/settings/<provider>.json templates with required isolation flags
• claude-profiles-init creates isolated ~/.claude-profiles/<provider>/ directories with symlinked shared resources
• Profile sync mirrors enabled plugins from the default Claude profile and shares installed plugin state
• Local source sync keeps Claude plugin caches source-backed, but links only the explicit Codex selection into ~/.agents/skills; a declared active subset may keep legacy ~/.codex/skills compatibility links
• A maintainer LaunchAgent watches the activation manifest and marketplace/install topology, while reporting non-compatible legacy links for reviewed cleanup instead of deleting them
• Built-in marketplace path pollution fixer runs automatically on every profile launch
• Includes student setup guide and troubleshooting reference

Example usage:

# Install the suite
claude plugin install daymade-claude-code@daymade-skills

# Then ask Claude naturally
"set up Claude Code profiles for Kimi and DeepSeek"
"I want to run Kimi and Anthropic in separate terminals"
"install the multi-provider profile setup from the workshop"

Requirements: claude CLI, zsh or bash, python3, plus API keys for the providers you want to use.

---

llm-eval-harness - Four-Dimension LLM Endpoint Evaluation

claude plugin install llm-eval-harness@daymade-skills

Evaluate any LLM behind an OpenAI- or Anthropic-compatible endpoint across four dimensions — instead of trusting a vendor's headline numbers — and report measured results separately from inferred ones.

When to use:

• Benchmarking a model, or verifying a vendor's tokens-per-second claim
• Comparing two models head-to-head under identical conditions
• Vetting a newly released or "Anthropic-compatible" endpoint before adopting it
• Probing concurrency limits before a workshop or batch job

Key features:

• Four dimensions, four scripts: speed (speed_probe.py — TTFT + thinking-aware tokens/sec), concurrency/stability (concurrency_probe.py — success rate, p50/p90, breaking point), Anthropic protocol compliance (protocol_probe.py — thinking-block trigger rate over N≥10), and quality regression (usecase_runner.py + independent blind judges)
• Thinking-aware throughput: captures reasoning_content separately so reasoning tokens don't inflate tok/s (the trap that once read a ~750 tok/s model as 4700)
• Probabilistic protocol verdicts: fully-implemented / intermittent (k/N) / not-implemented, never concluding from a single sample, with Connection: close so a load balancer can't pin all samples to one replica
• Blind-judge quality: 3 independent judges per case, majority-pass, with per-category precision to surface a systematically weak category
• Keys via env-var name only (--key-env MY_KEY) — the key never appears in ps, shell history, or a saved report
• Your use-case library lives outside the bundle (e.g. ~/.llm-eval/usecases.json) so it survives skill updates and never lands in a public repo

Example usage:

export MY_KEY=sk-...   # the key never appears in a command below

# Speed: real-task throughput + sustained decode ceiling
uv run --with openai python scripts/speed_probe.py \
  --base-url https://api.example.com/v1 --model some-model --key-env MY_KEY --mode both

# Concurrency: ramp until it breaks
uv run --with aiohttp python scripts/concurrency_probe.py \
  --url https://api.example.com/v1/chat/completions --model some-model \
  --key-env MY_KEY --format openai --concurrency 10 20 40 60

🎬 Live Demo

Coming soon

📚 Documentation: See llm-eval-harness/references/evaluation_disciplines.md for the reasoning behind each discipline and llm-eval-harness/references/quality_blind_judge.md for the blind-judge method.

Requirements: Python 3.8+, uv; openai and aiohttp (auto-installed via uv run --with); an API key for the endpoint under test. Optionally composes with promptfoo-evaluation for rubric-based gating.

---

read-claude-web-conversation - Extract Claude.ai Web Conversations

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:read-claude-web-conversation)

Extract the complete active Claude.ai conversation through the Claude Code operations suite: message text, real tool calls and outputs, uploads, generated images, and sandbox deliverables. Render faithful Markdown locally, reconstruct files created through tool calls, or download original bytes through the logged-in browser session.

When to use:

• Reading a Claude.ai web conversation that is not available in local Claude Code logs
• Recovering a full web thread for handoff, audit, or archival
• Comparing browser-visible content with exported or local session artifacts
• Preserving the analysis trace from agent conversations whose tool calls are collapsed in normal exports
• Downloading uploads and assistant-produced files that belong to the active conversation branch
• Using the macOS AppleScript channel when the Chrome extension is paired to a different account

Requirements: Installed daymade-claude-code suite and a logged-in Chrome session. The AppleScript fallback requires macOS and one user-enabled Chrome developer-menu toggle.

---

setup-notifications-via-wecom - Reusable WeCom Notification Setup

claude plugin install setup-notifications-via-wecom@daymade-skills

Set up reusable WeCom (Enterprise WeChat) webhook notifications for technical status reports, alerts, and completion messages. The target is explicitly classified: the user's own channel may send automatically; every other target requires human confirmation.

For recurring reports on an existing SLS pipeline, follow the bundled adapter guide for truncation limits, evaluation dates, recipient conflicts and native delivery acceptance.

When to use:

• Configuring a reusable 企业微信 / WeCom notification channel
• Sending structured status notifications, backup reports, or alerts
• Turning a one-off webhook into a repeatable notification workflow

Requirements: WeCom bot webhook URL and shell access.

---

notify-wecom - One-Off WeCom Message

claude plugin install notify-wecom@daymade-skills

Send a single WeCom group-bot message using the configured target identity: self sends directly, while others requires human confirmation.

When to use:

• /notify-wecom
• 临时发一条企业微信 / 企微通知一下
• One-shot alerts that do not need templates or persistent setup

Requirements: WeCom bot webhook URL plus the explicit recipient scope/label and canonical sender binding created by the setup command.

---

github-sensitive-data-cleanup - GitHub Sensitive Data Cleanup

claude plugin install github-sensitive-data-cleanup@daymade-skills

Scan and remove sensitive data from GitHub repository history, with backup, visibility checks, and force-push safety gates.

When to use:

• A repo leaked secrets, private domains/IPs, API keys, or PII
• Cleaning git history before or after a public exposure
• Verifying safety before any force push to a public repository

Requirements: git, GitHub access, and the relevant scanning/history-rewrite tools for the target repository.

---

codex-image-gallery - Local Browser for Codex Generated Images

Install: claude plugin install daymade-codex@daymade-skills (suite-only — invoked as daymade-codex:codex-image-gallery)

Start a self-contained local web gallery for Codex-generated image outputs. The skill bundles its Node server and HTML UI, scans ~/.codex/generated_images by default, and can point at another folder with GALLERY_ROOT.

When to use:

• Browsing Codex generated images in a local UI
• Inspecting ~/.codex/generated_images
• Reviewing a custom image output directory with search, batches, and image detail view

Key features:

• Bundled scripts/server.mjs and assets/index.html
• Dynamic /api/images scan; no hardcoded manifest
• /images/<relative-path> image route with path traversal protection
• Optional GALLERY_ROOT, PORT, and HOST

Requirements: Node.js 18+ and access to the image folder.

frontend-visual-qa - Rendered Frontend and Browser Output Visual QA Gate

claude plugin install frontend-visual-qa@daymade-skills

Audit the UI users can actually see, with explicit evidence levels and no source mutation by default.

When to use:

• Auditing an already-rendered web or desktop UI after implementation
• Diagnosing typography, wrapping, clipping, overflow, responsive, route/state, overlay, map, or transient-state defects
• Comparing a rendered artifact with a named visual reference or design-system SSOT
• Checking converted Markdown figures and illustrated transcript/note image sizes inside the actual Obsidian or other recipient reading view
• Verifying export, download, share, popup, print/PDF, or Electron-shell behavior at the evidence level the claim requires
• Complementing ui-designer/design work and the broader process managed by qa-expert

Key features:

• Audit-only default with scoped profiles and verified / partial / blocked outcomes
• A–D evidence ladder that prevents headless, renderer-only, or handler-only checks from masquerading as real GUI proof
• State/viewport-first workflow, including authenticated-but-role-less branches and exact projection/deck canvases, plus inspected screenshots and DOM geometry
• Hardened Playwright-powered sweep for navigation status, effective mobile viewport, overflow, clipping, custom-control focusability, images, and section evidence
• Conditional references for core visual checks, complex journeys/page contracts, and data visualization—including units/source/time/freshness, dense collisions, runtime-label truth, and real file-dialog boundaries
• Public behavior and trigger evals with self-contained fixtures

openclaw - OpenClaw (龙虾) Config Manager

claude plugin install openclaw@daymade-skills

Manage OpenClaw (龙虾/lobster) instance configurations — audit, diff, copy, add-model, list, and switch models across openclaw.json files.

When to use:

• Managing multiple OpenClaw / Claude Code wrapper instances
• Applying DeepSeek model patches to an instance
• Auditing, diffing, or copying provider/model config between instances
• Managing default models and aliases, or validating config

Key features:

• Unified CLI: audit / diff / copy / add-model / list / switch
• Auto-audit on mutating commands with a --no-audit escape hatch
• Nickname registry for cross-config operations

download-gemini-images - Download Images from Gemini Conversations

claude plugin install download-gemini-images@daymade-skills

Download images (uploaded files or generated previews) from a Google Gemini conversation page using your logged-in Chrome session, then package them into an ordered ZIP.

When to use:

• Saving images from a Gemini chat/app page (uploaded or generated previews)
• Need the larger lightbox image, not the thumbnail
• Renaming downloaded images in order and producing a ZIP archive

Key features:

• Lightbox-first download via the Chrome plugin (uses your existing Google session)
• pageAssets fallback when lightbox automation fails
• Ordered ZIP packaging with integrity verification

wps-doc-scraper - Archive Public WPS/KDocs Documents

claude plugin install wps-doc-scraper@daymade-skills

Faithfully archive public WPS / KDocs / 金山文档 links — especially embedded ProcessOn mind maps and canvases — as raw source data, original SVG/PNG, and Markdown, without logging in.

When to use:

• Given a kdocs.cn or wps.processon.com link to scrape, save, download, or convert to Markdown
• Archiving an embedded ProcessOn mind map or canvas with source fidelity
• Need the raw payloads + original visual artifact, not just rendered text

Key features:

• Unauthenticated data-API-first extraction (no login, no account save)
• Original SVG/PNG capture for canvases and mind maps
• Markdown as a structured representation of the source

ashare-news-fetcher - A-Share Market News & Sentiment Aggregator

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:ashare-news-fetcher)

Aggregate A-share (Chinese stock market) news, policy, and sentiment from public sources — 财联社, 华尔街见闻, 金十, 新浪 7x24, 东财快讯, regulator announcements (CSRC / PBoC / SSE), 东方财富股吧 — into structured JSON or Markdown.

When to use:

• Need recent news/policy/sentiment for a specific A-share stock or the whole market
• Aggregating 消息面 intel from multiple Chinese financial sources at once
• Producing a structured JSON or Markdown news digest

Key features:

• Multi-source public-feed aggregation (no login)
• Per-stock or market-wide filtering
• Structured JSON / Markdown output

local-codex - Local OpenAI Codex CLI Agent

Install: claude plugin install daymade-codex@daymade-skills (suite-only — invoked as daymade-codex:local-codex)

Delegate coding tasks to the local OpenAI Codex CLI agent using your ChatGPT Pro OAuth flat-rate subscription. Wraps codex exec / codex review for code generation, refactoring, and review without per-token API charges.

When to use:

• Delegating coding tasks to Codex for generation, refactoring, or review
• Running codex exec for non-interactive coding tasks
• Using Codex's GPT-5.5 agent capabilities through local CLI

Key features:

• ChatGPT Pro OAuth flat-rate billing via ~/.codex/auth.json (no API keys)
• Auto-detects Codex CLI from desktop app, Homebrew, npm, or PATH
• JSONL output parsing with final assistant message extraction
• Configurable sandbox levels: read-only / workspace-write / danger-full-access

openclaw-model-switch - OpenClaw Model Switcher

claude plugin install openclaw-model-switch@daymade-skills

Switch the default AI model for an OpenClaw instance (e.g., Kimi K2.6 → K2.7) by safely editing openclaw.json with automatic backup, model validation, and optional gateway restart.

When to use:

• Upgrading to a newly released model
• Switching between models for different tasks
• Rolling back after testing a new model

Key features:

• Automatic config backup before changes
• Model validation against provider's model list
• Optional gateway restart after switching
• Supports Kimi K2.x model family with context window specs

pharma-daily-report - A-Share Pharma Sector Daily Report

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:pharma-daily-report)

Generate an A-share pharmaceutical sector daily report — pull real-time quotes for core pharma stocks from Sina Finance, rank 7 sub-sectors, top gainers/losers, and estimated fund flow, then optionally push a rich-text report via Feishu.

When to use:

• Need a daily snapshot of the A-share pharma sector (quotes, sub-sector ranking, fund flow)
• Want a Feishu rich-text push of pharma market data
• Tracking a configurable watchlist of core pharma stocks

Key features:

• Sina Finance real-time quote pipeline (no login)
• 7 sub-sector classification + gainers/losers + fund-flow estimate
• Optional Feishu rich-text delivery; default 20-stock watchlist, customizable

gemini-history-analyzer - Analyze Gemini Conversation History

Install: claude plugin install gemini-history-analyzer@daymade-skills

Analyze Google Takeout exports of Gemini conversation history — extract and categorize transcripts and attachments, mine them for domain-specific insights with context-verified keyword search, and optionally distill findings into a personal knowledge base.

When to use:

• Have a Gemini Takeout zip and want to know what's in it (topics, conversation type, valuable documents)
• Want to mine Gemini history for a specific domain (finance, legal, etc.) without drowning in keyword false positives
• Distilling Gemini conversation data into project memory or a personal wiki

Key features:

• Handles Chinese/Unicode filenames (uses unar, not the corrupting macOS unzip)
• Meeting-transcript vs prompt-response detection; topic categorization; PII flagging
• Context-verified keyword search (grep is step 1, not the answer) + optional memory-file generation

skill-governance - Real Skill-Surface Governance

Install: claude plugin install daymade-skill@daymade-skills (suite-only — invoked as daymade-skill:skill-governance)

Govern the Skill surface Claude Code and Codex actually expose without losing cold capability. It separates canonical source, installed inventory, discovery policy, the fresh model-visible catalog, and router-resolved runtime resources.

When to use:

• Codex reports many Skills, truncated descriptions, duplicate identities, or a missing router
• A Skill appears stale, missing, duplicated, or installed from an unexpected source/version/scope
• Need to reconcile owned source links under ~/.agents/skills without treating every disk bundle as active
• A merged suite migration left old standalone plugin identities installed locally
• A third-party bundle must stay on disk for runtime resources while only its router remains visible to Codex

Key features:

• Compares the actual fresh Codex prompt with complete metadata from Codex's own skills/list, detecting truncation, omitted enabled entries, identity/source collisions, disabled-path drift, scan errors, and missing direct activation entries
• Treats owned source as canonical, plugin caches as derived runtime state, and preserves current install scope
• Routes source-backed activation to its explicit manifest owner instead of mixing third-party inventory into it
• Discovers suites dynamically from the manifest and verifies the replacement suite before retiring standalone installs at their original scopes
• Keeps third-party bundles as cold disk inventory through exact-path Codex discovery policy, then verifies both the clean prompt catalog and preserved runtime resources
• Leaves Claude's versioned orphan-cache lifecycle to Claude; manual cache removal is exceptional, recoverable repair rather than routine cleanup

photo-to-scanned-pdf - Phone Photos to Scanner-Quality PDF

Install: claude plugin install daymade-docs@daymade-skills (suite-only; invoke as daymade-docs:photo-to-scanned-pdf)

Turn phone photos of contracts, receipts, forms, certificates, or handwritten pages into a clean A4 PDF. The workflow combines perspective correction, scanner-style background cleanup, colored-paper handling, explicit content-based page ordering, and a mandatory whole-document contact-sheet check.

github-review-pr - Current-Base Contributor PR Review

claude plugin install github-review-pr@daymade-skills

Review or re-review one named contributor PR—including a closed PR under reconsideration—or sweep every open PR newest-to-oldest as a repository maintainer, always against the live base rather than a stale PR snapshot.

When to use:

• Main/base changed after an earlier review and the PR needs a fresh decision
• A specifically named closed PR needs a merit review or reconsideration
• Need to distinguish PR-owned defects from pre-existing base debt
• Need to know what the actual three-way merge would land now
• Need a newest-to-oldest decision ledger for the complete open-PR queue
• Need a review-gated repair or merge, with every GitHub write explicitly authorized
• Explicitly want an independent maintainer-style evaluation of your own contribution; this remains a read-only recommendation, with no self-approval or merge authority
• Want the opt-in personal policy to learn from prior closed-PR decisions, enforce the maintainer's curation bar, or decide whether worthy original contributor PRs should be repaired instead of recreated

Key features:

• Separates PR-recorded base, current base, and head OIDs, then rechecks them before the verdict
• Reviews the prospective merge tree when clean, or isolated intent plus conflict stages when no landing tree exists
• Detects ordinary base drift versus history discontinuity, then separates polluted branch history from the contributor's verified intended patch
• Can project that verified patch onto current base for analysis while labeling the result synthetic and non-landable
• Requires target code, tests, runtime behavior, logs, or repository specification for findings; reviewer concurrence raises confidence, not severity
• Attributes every finding as PR, BASE, or SHARED before assigning repair ownership
• Distinguishes a curation DECLINE from a fix request or true supersession
• Keeps review read-only by default; repair, review posts, close, merge, admin, auto-merge, and branch deletion remain separate actions
• Treats external-contributor program goals only as a post-merit priority, never a reason to lower the acceptance bar
• Requires one fresh, context-bound confirmation for the single surfaced PR before merge; a short affirmative reply is accepted without carrying authority to the next PR

Example usage:

/github-review-pr --personal-maintainer https://github.com/<owner>/<repo>/pull/<number>
/github-review-pr --personal-maintainer --all-open
main changed since the last review; tell me what would actually land now
review all of our open PRs newest to oldest and apply my prior maintainer principles
can we merge this contribution and fix the remaining repo bookkeeping ourselves?

Requirements: authenticated gh CLI, git with merge-tree --write-tree, and jq.

read-codex-history - Read Local Codex History

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:read-codex-history)

Read, search, and export local Codex history without continuing the old task. It keeps the prompt ledger, state database, and rollout JSONL as three distinct evidence surfaces so exact user input, inventory metadata, and Agent behavior are never silently substituted for one another.

Key features:

• Lists Codex sessions with internal time ranges and active/archive provenance
• Extracts exact prompt-ledger inputs newest-first and groups them only by Session
• Reads bounded current-thread inputs appended after a fixed canonical-source cursor; returns original input, never an approval
• Reconciles whole-conversation input counts and literal quotations across exact inherited snapshots; reports unresolved membership and accepts only record-bound reviewed injection exclusions
• Reconstructs one rollout as a chronological user/assistant timeline with exact fork byte boundaries and compacted context
• Searches Codex rollouts only; it cannot silently mix Claude matches into a Codex request
• Selects a compatible Codex state database through schema introspection
• Visibly falls back to raw Codex rollout JSONL when the database is unavailable
• Requires internal session_meta.id equality and reports fused/missing rollout gaps instead of guessing
• Probes every in-scope Codex session and appends any row whose exact canonical writer-lock file is held even beyond the recent-row limit; the marker proves lock state, not holder identity or a running agent, and an unmarked row is never reported as stopped or safe to retire
• Supports internal-time date windows, custom exact-scope diagnostics, archived Codex threads, recursive/all-project scopes, Windows path normalization, and JSON
• Uses Python's standard library only; performs no network requests or writes

Example usage:

/daymade-claude-code:read-codex-history
list the recent Codex chats for this folder
show my 200 most recent original Codex inputs, grouped by Session
read Codex Session 01abc... chronologically and show its fork lineage
show which Codex sessions in this workspace have held canonical writer-lock files
show Codex threads including archived conversations as JSON

📚 Documentation: See storage_and_portability.md for source selection, path normalization, privacy boundaries, and diagnostics.

Requirements: Python 3.10+; no third-party packages or network access.

---

continue-codex-work - Resume Interrupted Codex Work

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:continue-codex-work)

Continue a prior Codex task without replaying it through codex resume. The Skill first requires read-codex-history to prove selected identity, fork lineage, chronology, compacted context, and gaps; it then rebuilds the original business outcome, unfulfilled requests, user corrections, proven assets, and one direct next action before changing the project. It is for a new/different Agent context taking over an earlier rollout. When Codex itself natively resumes the same conversation and prior context is already present, continue directly instead of invoking this Skill.

/daymade-claude-code:continue-codex-work 019f66...
Codex was cut off mid-task; recover that rollout and finish the work

📚 Documentation: See continue-codex-work/SKILL.md.

Requirements: Python 3.10+ and local Codex rollout files.

---

git-safety-net - Prevent & Recover From Local-Git Disasters

Install: claude plugin install git-safety-net@daymade-skills

Prevent and recover from the branch/stash/rebase tangles that strand or lose local commits, detached or dirty worktrees, and answer "is everything actually merged?" with content-level proof instead of misleading commit counts. Every command is non-destructive or additive until a step is explicitly labeled destructive.

Key features:

• Recovers deleted commits/branches/stashes via git reflog and git fsck (the ~90-day window)
• Audits every linked worktree HEAD, dirty checkout, local-only commit, stash, and dangling commit
• Pins dangling commits gc-proof under refs/dangling-backup/ before any cleanup
• Verifies a branch is merged by CONTENT — defeating the squash-merge "100 commits ahead" illusion
• Proves a linked worktree safe before non-forced removal and exports all refs in one verified bundle
• Optional adversarial multi-agent verification for a high-stakes "is everything merged?" call
• Catches another session's commit adopted onto your branch before it ships inside your PR
• Settles "did my push/merge actually land?" by content when the receipt was lost to a flaky network
• Prevention habits: commit before switching, push WIP early, audit every checkout before deletion

Example usage:

/git-safety-net
did I lose any commits after all that branch switching?
is everything merged into main, or is something still stranded?
recover the commit I lost after a bad rebase
prove this worktree can be deleted without losing anything

📚 Documentation: See recovery_playbook.md, merge_verification.md, and prevention_practices.md.

Requirements: Git and a standard Bash shell. No third-party packages; current remote status requires network access, while offline audits use cached remote-tracking refs.

---

design-style-picker - Batch-Compare Visual Design Directions

Install: claude plugin install daymade-codex@daymade-skills (suite-only — invoked as daymade-codex:design-style-picker)

Turn vague taste into concrete visual choices. Instead of guessing one final design, batch-generate a structured set of visual directions so the user picks the style they actually want — the goal is exposing the taste boundary quickly, not reading minds.

Key features:

• Batch-generates and compares multiple design directions side by side
• For users who cannot describe an abstract visual style in words
• Works from generated UI / design-system images; rejects "too colorful / too dead / too generic" with more options
• Evolves an existing UI or design system without discarding current assets
• Ships a selection playbook and reusable prompt patterns

Example usage:

/daymade-codex:design-style-picker
I can't describe the style I want — show me a batch of options
this draft is too colorful and too generic, generate more directions
evolve our current design system without throwing away existing assets

📚 Documentation: See selection-playbook.md.

---

claude-migrate-memory-to-doc - Migrate Claude Memory to Tool-Agnostic Docs

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-migrate-memory-to-doc)

Migrate Claude Code personal memory (the per-project memory/ directory) into tool-agnostic reference docs, so other AI CLIs that auto-load AGENTS.md (Codex primarily; the architecture transfers to Cursor and others) working in the same directory read the same user profile and collaboration preferences instead of being blind to them.

Key features:

• Solves tool lock-in: memory is not version-controlled and is invisible to every non-Claude tool
• Two-layer architecture: references/ + CLAUDE.md-inline + AGENTS.md-symlink, designed around what each tool actually auto-loads (plain-text pointers are on-demand in both tools)
• Full workflow: diagnosis, multi-agent review, empirical codex verification, and memory cleanup
• Two goals: end tool lock-in, where memory stays as a thin handoff cache; or retire a project's auto memory, where every entry moves to a document, the project switch goes off, and a control probe confirms memory no longer loads
• Runs inline, orchestrating review subagents and invoking codex directly

Example usage:

/daymade-claude-code:claude-migrate-memory-to-doc
migrate my memory — Codex doesn't know who I am
my memory is locked to Claude Code, make it tool-agnostic
memory has grown bloated with content that should live in docs
turn off auto memory for this project and move everything into docs

---

docx-creator - Produce Production-Grade Word Documents

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:docx-creator)

A thin incremental layer over minimax-skills:minimax-docx — not a document engine itself. Adds the layer the underlying OpenXML engine doesn't ship: a verified markdown-to-docx generator for Chinese formal documents, the alignment-layering rule that stops justified text from stretching 甲方/乙方 info blocks and signature blocks into garbage, per-list numbering restart, CJK font dual-slot setup, and a mandatory LibreOffice-to-PDF-to-PNG visual verification chain (qlmanage thumbnails are banned — they hide exactly the bugs that matter).

Key features:

• Division of labor: the OpenXML SDK engine stays in minimax-docx; this skill owns Chinese formal-document typography rules and the CLI-vs-C# routing decision
• Verified markdown-to-docx generator (scripts/Program.cs) for contracts, agreements, and 公文 with 甲乙方 info blocks, numbered clauses, signature blocks, and tables
• Mandatory real-render visual verification (LibreOffice → PDF → PNG), not a Quick Look thumbnail
• Routes plain prose to the minimax-docx CLI instead of duplicating it

Example usage:

/daymade-docs:docx-creator
生成 Word 文档
写合同 docx
把 markdown 转成 Word
give me a Word file for this labor contract

---

read-docx-review - Read Word/WPS Review Comments & Tracked Changes

Install: claude plugin install daymade-docs@daymade-skills (suite-only — invoked as daymade-docs:read-docx-review)

Read a reviewed docx (Word/WPS) and extract every comment and tracked change into a per-item adjudication table (markdown or JSON): who commented, on which paragraph, what they said, plus revision-mode insertions and deletions. Revision-aware engine on the OpenXML SDK, so content inserted via track changes (w:ins) and paragraph-level deletions (w:del) that a bare python-docx read silently misses are captured. Read-only — never mutates the reviewer's file.

Key features:

• Per-item adjudication table with a blank disposition column, built for clause-by-clause triage
• Captures the w:ins / w:del revisions that python-docx silently drops
• WPS reference-only comment anchors get a fallback so every comment stays locatable
• Zero-comment files emit an explicit "no review traces" warning instead of a silent empty table
• Boundary: producing/formatting docx → docx-creator or minimax-docx; PDF annotations out of scope

Example usage:

提取这份 docx 的批注
对方批注完的合同回来了，读一下审阅意见
读一下修订，谁批了什么、批在哪

---

claude-code-hooks - Write, Test, and Debug Claude Code Hooks

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-code-hooks)

How to write, test, register, and debug Claude Code hooks — PreToolUse / PostToolUse / SessionStart / Stop Bash guards that enforce a rule the model would otherwise talk itself past. A prose rule in CLAUDE.md is a suggestion the completion drive can override; a hook is a wall.

Key features:

• Five runnable pattern skeletons — PreToolUse block, human-confirmation release gate, SessionStart health check, PostToolUse context injection, and a Stop hook that reacts to the model's own output — plus the shlex command-position walker for token-level command matching
• Hard-won rules, among them: shlex over awk-splitting so a healthy command is never false-blocked, bash -n + real-JSON end-to-end testing before registering, version-controlled source with installer-owned recovery, and per-profile registration convergence with human-confirmation release gates
• Cataloged failure modes with symptom → cause → fix, including the UserPromptSubmit-vs-Stop category mistake (only Stop can see what the model itself wrote) and a literal quote/backtick inside a Python comment silently corrupting an embedded python3 -c block
• Bundled end-to-end test harness (scripts/test_hook.sh)

Example usage:

/daymade-claude-code:claude-code-hooks
make a hook that blocks git push --force
turn this rule I keep forgetting into a hard gate
my hook is false-blocking a healthy command
debug this hook that's poisoning my session

macos-watchdog - Design & Discipline macOS launchd Watchdogs

Install: claude plugin install daymade-macos@daymade-skills (suite-only — invoked as daymade-macos:macos-watchdog)

Design, deploy, and discipline macOS launchd watchdogs — LaunchAgents / LaunchDaemons that detect a recurring problem and auto-remediate it, without the watchdog itself becoming the disturbance.

📚 Documentation: macOS Watchdog; for persistent Python deployment, follow its runtime procedure.

Example usage:

make a launchd watchdog that reconnects my VPN when it drops
my watchdog keeps re-launching an app I quit — make it stop
audit my existing LaunchAgents for noise and crash loops

macos-load-doctor - Diagnose macOS System-Level Load and Process Leaks

Install: claude plugin install daymade-macos@daymade-skills (suite-only — invoked as daymade-macos:macos-load-doctor)

Find what is making the Mac slow or hot, attribute it to an owner, and act within the shared-machine boundary. Load average first, then a three-reading process census, parent-chain attribution, and shape classification — the skill's job is a correct, evidence-backed attribution, not process cleanup.

Key features:

• The fleet-wide rule — when independent apps/hooks/agents all time out at once, the machine is the suspect, not any of them: vm.loadavg before any individual debugging
• Three-reading census (bundled load_census.sh) — children-per-parent (the leak that no orphan reaper will ever clean), cumulative CPU time (the busy loop %CPU dilutes), instantaneous %CPU (the active storm)
• Shape classification that picks the right fix — leak / fork storm / busy loop / cascade — with a hard shared-machine boundary: agents diagnose read-only and report; the owner terminates (a "daemon restart" counts as terminating)
• Incident playbook with worked cases (per-thread MCP spawner leak at load 243, unthrottled replay fork storm, guard-fleet fork overhead, GUI busy loop) and the report-writing contract

Example usage:

everything on this Mac is slow and every agent's hooks are timing out
the machine is burning hot and I can't tell why
hundreds of node processes — where are they coming from?

devils-advocate - Pressure-Test an Investment Thesis

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:devils-advocate)

Structured devil's-advocate pressure-testing of an investment thesis against user-supplied evidence materials — evidence-anchored dissent, not role-played contrarianism (every counterpoint must cite, verbatim).

Key features:

• Decomposes the thesis into explicit assertions and implicit assumptions (fact/forecast/mechanism typing, load-bearing test)
• Retrieves counter-evidence per assumption under a source-credibility ladder with verbatim citations, plus a Mauboussin base-rate outside view bounded to your materials
• Emits an auditable findings JSON and a theme-grouped analyst narrative, and converts the critique into a RAND-style signpost monitoring list

Example usage:

/daymade-financial:devils-advocate
stress-test my bull case on <stock> against these three research reports
what would have to be true for this thesis to be wrong?

daymade-sector-research - A-Share Sector Research Workflow

Install: claude plugin install daymade-financial@daymade-skills (suite-only — invoked as daymade-financial:daymade-sector-research)

A-share (Chinese market) sector research workflow: Top-N gainers across all sector constituents, announcement-window retrieval, and market-sentiment judgment with graded evidence — executed by an Agent Team with fresh-context adversarial verification. All public, no-login data sources.

Key features:

• Top-N gainer pipeline across full sector constituent lists (East Money push2 + Sina realtime snapshots, cross-checked)
• Announcement windows per stock (weekly/monthly) from cninfo + East Money, covering Shanghai/Shenzhen/Beijing exchanges
• Sentiment verdicts graded by evidence level — L1 first-hand quotes / L2 timestamped media / L3 unverified headlines — with "rather uncertain than wrong" as the standing discipline

Example usage:

/daymade-financial:daymade-sector-research
医药行业今天的 Top 10 标的
这些标的最近一个月发过哪些公告
判断医药板块现在的市场情绪

kimi-use - Kimi Desktop Plugins and Deep Research

Drive the logged-in Kimi desktop app through computer-use to query its built-in company, financial, market, academic, and legal data plugins without separate API keys. Results stay source-labeled and must be independently checked before they become load-bearing data. Its separate Chat Deep Research route exports a full report and attachments.

Key features:

• Routes requests to installed Kimi plugins such as 天眼查, 同花顺 iFinD, SEC, IMF, and academic/legal databases
• Covers both Claude Code computer-use and Codex computer control
• Provides query patterns that require explicit sources and honest unknowns
• Guards against truncated lists, OCR/name errors, and financial-data scope mismatches
• Separates Work/Agent plugin calls from Chat Deep Research and preserves the original report export
• Checks per-interface credits before autonomous Work tasks and retains distinct raw returns for retries

Example usage:

/kimi-use
用 Kimi 查这家公司的股东和最新财务数据，并逐项标明来源
操作 Kimi 客户端，用已安装插件核对这组市场数据

tibo-reset-codex - ChatGPT/Codex 额度重置速查

Install: claude plugin install tibo-reset-codex@daymade-skills

查询重置公告，核实多个 Pro 账号的剩余额度与备用重置；有时间依据时预测下一轮，记录预测、核验与撤回，供后续判断调整。

操作说明

ChatGPT 什么时候重置额度
几个 Pro 账号都用完了吗，哪个还有额度
banked reset 到了吗

prior-work-retrieval - Retrieve Proven Work Before Producing

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:prior-work-retrieval)

When the user explicitly asks to reuse or find earlier work, verify successful code, decisions, Skills/SOPs, meetings, WeChat archives, documents, and conversation history before producing. References to the current tests, README, files, implementation, or behavior do not trigger retrieval. Every real run leaves an auditable reuse/adapt/reject receipt; zero ranked hits never become an absence claim.

Key features:

• Explicit source manifest and per-carrier coverage
• Original-source authority and freshness verification
• Reuse/adapt/reject decisions tied to the current business outcome
• Deterministic receipt check before substantial production

Example usage:

/daymade-claude-code:prior-work-retrieval
we solved this before; find the code and successful path before changing anything
don't rebuild the workflow until you have checked our existing Skills and history

lark-cli-router - One Version-Synced Lark Router

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:lark-cli-router)

Route Feishu/Lark/Doubao requests to the matching guide embedded in the installed lark-cli, instead of loading every Lark domain Skill into the model catalog. Tencent IMA stays separate under ima-skill.

Key features:

• Selects the smallest matching domain or workflow guide from lark-cli skills list
• Reads version-matched instructions and references through lark-cli skills read
• Keeps unembedded scripts/assets reachable with an embedded-vs-disk guide hash gate
• Preserves each domain's auth, confirmation, success, and verification contract

Example usage:

读取这个飞书文档
查一下这条妙记的逐字稿
lark-cli says the user identity is missing a scope

claude-code-ping-start-5h-quota - Ping Claude After Quota Reset to Start a Fresh 5h Window

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:claude-code-ping-start-5h-quota)

Schedule a one-shot local timer (macOS) that sends one minimal claude -p message right after your subscription quota resets, so the next 5-hour usage window starts counting while you sleep instead of when you wake up. Dual redundancy: the detached ping survives closing the session, and a live session is re-woken to verify the ok receipt.

Key features:

• Cross-midnight absolute-time handling and an explicit minutes→seconds formula with a 5-minute buffer
• caffeinate -is holds off idle sleep during the wait; independent ps read-back verifies the timer chain
• Persists the trigger time and output path into the conversation, so a lost wake-up notification stays recoverable
• Feature-string pkill cancel command (live-verified) — never kill by sleep seconds
• Honest failure reporting: says "cannot confirm locally" instead of guessing

Example usage:

还有 1 小时 20 分钟重置额度，帮我定时戳一下 Claude
我去睡觉了，额度重置后帮我 ping 一下开启新窗口

agent-web-search-setup - Working Web Search When the Model Backend Cannot Run It

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:agent-web-search-setup)

A reseller or relay that proxies Claude or Codex to another cloud (Vertex AI, Bedrock, or an OpenAI-compatible layer) cannot execute Anthropic's server-side web_search and web_fetch. The request still succeeds, so nothing errors — the model simply gets no results and reports that recent things do not exist, which reads as "the model got stupid" rather than as a broken tool. This skill probes the endpoint to establish which built-in tools are actually dead, removes those from the client so the model stops reaching for them, installs a client-executed replacement it can really call, and proves it with a live query. It configures the agent; it is not itself a search engine.

Key features:

• diagnose.py (standard library only, runs before anything is installed) reads the shape of the reply rather than its text: server_tool_use means the tool really ran, a plain tool_use means it was handed back and never executed, HTTP 400 means the translation layer cannot express the tool at all
• Covers Codex as well as Claude Code: --api openai asks the same question of the OpenAI Responses API, where a working endpoint returns a web_search_call item, and --api both handles a relay serving both shapes
• Most "give your agent internet access" bundles orchestrate the tools the model already declares, so on a broken relay they inherit the exact failure and hide it behind a convincing layer; the skill says how to tell those apart from the ones that shell out to a local process
• The registered CI suite is weighted toward the expensive direction — calling a healthy endpoint broken. A text-only reply, an empty reply, a timeout and a missing route each have a case demanding inconclusive, because every one of them would otherwise have the caller delete a working tool
• The grant is where this goes wrong in practice, so it is pinned by a test: permissions.allow in a project's shared .claude/settings.json is dropped in any workspace whose trust dialog has not been accepted, and the client says so on stderr alone — so an agent reading --output-format json sees a refusal with no reason, and the model tells the user it is waiting for permission
• Names the real backend from the tool-call id prefix (toolu_vrtx_ Vertex AI, toolu_bdrk_ Bedrock), which is why switching models on the same endpoint does not help
• Removing the dead tool is a precondition, not a preference: with WebSearch still on the menu, a model made 21 calls to it and 0 to an installed alternative whose description said the built-in was broken
• A replacement menu with one default that needs no account and no local runtime, live-tested alternatives kept as real options, and the rejected candidates each carrying its reason
• Slow is never reported as dead — one gateway was measured answering in about 15 seconds on two model ids and about 146 on a third, so a timeout returns inconclusive
• Verification runs with only the newly installed backend available, because a second working one will answer for a broken one and turn the check green on a configuration that does not work

Example usage:

搜索返回空的，模型说这东西不存在
我用的是中转站的 key，联网搜索好像坏了
give this agent internet access — web search does nothing

tech-selection - Gated Checklist for Choosing Technologies

Install: claude plugin install daymade-claude-code@daymade-skills (suite-only — invoked as daymade-claude-code:tech-selection)

A checklist for choosing between technologies — library, framework, storage, data format, model, build-vs-buy, architecture. Runs before committing, not after. The criteria are filters, not sorters: they kill candidates that violate a principle, and survivors are decided by business-result anchoring, never by ranking. When two or more candidates survive, the output is candidates

• trade-offs + a recommendation, never a single pick.

Key features:

• Triggers on scenario sentences (用哪个 / 选什么框架 / 要不要自建 / 先看看有没有现成的) and characteristic negations (别闭门造车 / 不要过度工程), not the literal term 技术选型 which the user's own corpus almost never contains
• unknown is not pass — a candidate carrying an unverified axis does not enter the survivor set
• Two stops that return to the user: multi-candidate human tradeoff (≥2 survivors) and unverified completion claim
• Prior-art inventory as step 1 — internal/paid assets → external solutions → build from scratch (last resort); recommending build without recorded reasons from layers 1–2 is flagged as 闭门造车
• Probe-based evidence only — READMEs, vendor pages, and source-code claims are downgraded; the only admissible evidence is behavior you ran and observed
• Self-defense slot ("why this isn't garbage") that must be independently checkable, not self-certified
• Delegation contract with domain ownership table, three-part autonomy threshold, and 5 resolved scope boundaries
• Agent orchestration by task shape (four-question framework), not a fixed default
• 13 core decision axes + 13 scoped criteria + 16 rejection patterns, each with a mechanical test rather than an attitude
• 4 references: decision-axes, scoped-criteria, rejection-modes, delegation-contract

Example usage:

用哪个框架？React 还是 Vue？
这个爬虫用 requests 还是 playwright
数据存哪里？SQLite 还是 Postgres？
要不要自建，还是用现成的？
别闭门造车，先看看有没有现成的
which library should we use for this?
build or buy — evaluate the options

codex-1m-context-window-setup - Model-Aware Long Context for Codex

Install: claude plugin install daymade-codex@daymade-skills (suite-only — invoked as daymade-codex:codex-1m-context-window-setup)

Configure the shared Codex CLI/Desktop base config for the largest context window the selected model actually declares, up to a one-million-token request. The Skill discovers the live model contract, explains the familiar ~258K default, applies a 60% auto-compaction threshold, preserves unrelated TOML, backs up changed bytes, and rolls back if Codex strict-config validation fails.

Key features:

• Model-aware ceiling instead of a hard-coded 1M claim
• Read-only doctor, transactional apply, and drift-detecting verify modes
• Exact rollback, idempotent no-op behavior, and macOS/Windows support
• Changes only model_context_window and model_auto_compact_token_limit

Example usage:

/daymade-codex:codex-1m-context-window-setup doctor
configure Codex CLI and Desktop for the largest verified context window
verify that my long-context setup still matches the selected model

---

interaction-design-board - Runnable Interaction Alternatives Before Implementation

Install: claude plugin install daymade-codex@daymade-skills (suite-only — invoked as daymade-codex:interaction-design-board)

Generate several genuinely different HTML interaction architectures for one bounded product surface while holding business facts and the current design language constant. The bundled builder combines the self-contained candidates into a keyboard-accessible Design Board with Focus/Compare modes, structured task feedback, remix requests, and a content-bound approval handoff. Product code stays untouched until the user has operated and approved a direction.

Key features:

• Distinguishes interaction architecture from static visual-style exploration
• Keeps one frozen business/task contract across isolated candidate generation
• Rejects byte-identical candidates, path traversal, missing states, and external assets
• Supports gstack's local feedback daemon and direct-file JSON download fallback
• Preserves prototype content identity separately from the runtime Board instance

Example usage:

/daymade-codex:interaction-design-board
show me three clickable ways to organize this review workflow before changing the app
the visual style is fine, but compare queue-first, object-first, and ledger-first interactions
use a Design Board so I can try each progressive-disclosure option and request a remix

📚 Documentation: See interaction-design-method.md and board-contract.md.

---

meme-creator - Video/GIF Memes with Motion-Tracked Overlays

Glue logos, avatars, or stickers onto moving objects in a video clip so they follow the motion frame by frame — then export MP4 + GIF. Semi-supervised tracking: you read boxes off grid sheets, OpenCV CSRT carries them, and re-anchors or hand-set keyframes take over where trackers die (shot changes, walk-toward-camera scale blowups, long smooth walkaways).

When to use:

• Making a meme, 梗图, or reaction GIF out of an existing video moment
• Covering faces/heads in a clip, or putting a logo/avatar on a moving person or object
• "贴 logo 到视频里跟着动" — any overlay that must track motion

Key features:

• Identity disambiguation gate before binding any name to an account/avatar/logo (enumerate candidates with a handle-free search; the peer entities in the request are the discriminating signal)
• Segment picking from tiled contact sheets instead of scrubbing
• CSRT tracking with backward tracking, segment re-anchors, and smoothing; manual keyframes with piecewise-linear interpolation for smooth long shots
• Visibility windows and velocity-extrapolated fades, so badges leave the frame with their subject instead of parking mid-screen
• Two-pass palette GIF encoding with a size-budget knob order (fps → width → colors)

Example usage:

"把这段视频里三只猫的头分别换成这三个 logo，做成梗图视频和 GIF"
"Cover the CEO's face with our competitor's logo in this keynote clip"
"Turn 6:30-6:50 of this bilibili video into a GIF with my avatar on the main character"

📚 Documentation: See meme-creator/SKILL.md and the bundled references/ for the tracking playbook and the asset-binding gate.

Requirements: ffmpeg; uv (bundled Python scripts carry inline dependencies). yt-dlp only when downloading from a URL.

---

data-visualization-discipline - Judgment Layer for Charts and Dashboards

Install: claude plugin install data-visualization-discipline@daymade-skills

Decides whether a chart should exist and how it should be drawn before any code is written: the one conclusion each chart must support, mean vs. median, whether segments may be aggregated, stacked vs. line, dual axes, color count and cross-chart color identity, and how many columns a table keeps. Medium-agnostic — HTML reports, React/Vue dashboards, native PowerPoint charts, matplotlib/plotly/ECharts/D3.

Key features:

• Five stages (intent → data validity → form → encoding → delivery): the first four end with a handoff check, the fifth is the delivery gate

• A nine-item delivery gate, including a masked-render self-test and painted-span proportion checks

• Reviews existing charts ("what is wrong with these charts", "why do the legends disagree")

• data-visualization-discipline — full instructions. Pairs with report-with-html, which owns the report page itself.

---

report-with-html - Evidence-Backed HTML Reports

Install: claude plugin install report-with-html@daymade-skills

Produces the reader-facing HTML artifact — report, dashboard, architecture or journey view, data browser, or review workbench — built around the reader's question, with every number traceable and the page verified in a real browser before delivery.

Key features:

• Skeleton chosen by the reader's question (how it works / what to decide / why it happened / which is better / inspect a collection)

• Decision-card contract, reusable interaction components, and a warm-paper starter template

• Delivery gate script with a masked (text-free) render, segment cropping for full-page review, and a zero-context independent reader review

• report-with-html — full instructions. Load data-visualization-discipline for chart judgment.

Requirements: Google Chrome or Chromium; uv; Python 3.10+.

---

skill-evolver - Research-Driven Skill Improvement

Install: claude plugin install skill-evolver@daymade-skills

Daymade Skill Evolver assembles an improvement workflow for a specified existing Skill, using existing Skills for evidence gathering, mechanism trade-offs and authorized implementation. Reuse successful and failed work, inspect relevant competitor source, and verify changes through the target's actual consumer. Keep the current design when the evidence supports no change.

When to use: Improve a named Skill from work history, saved research or competitor mechanisms. Use Skill Creator for routine creation and editing.

Example usage:

Use skill-evolver to improve my batch-export Skill. Preserve successful exports,
investigate the interrupted-batch failure, and implement the justified repair.

Requirements: A host with access to the target source and the tools needed to edit and verify it. Resolve, load and execute available public stage Skills; local private integrations are optional. Missing required sources or data pause only dependent work.

Dependencies: This plugin ships only the orchestrator. For a missing executor, report its supported install package from the public catalog or an explicit fallback with its evidence limits.

📚 Documentation: skill-evolver.

---

app-evolver - Evidence-Based Application Improvement

Install: claude plugin install app-evolver@daymade-skills

Improve an existing application from its current code and product contract, authorized successful/failed work, and competitor mechanisms. Decide what fits, implement the authorized increment and verify the actual user result.

When to use: Take the useful parts of a competitor, improve an app from past experience, or coordinate improvements across apps with separate contracts. Use skill-evolver when the target is a Skill.

Use app-evolver to improve my document app from this competitor. Preserve source
provenance, implement the justified improvement and verify failure recovery.

📚 Documentation: app-evolver.

Reverse Skill composition: Specialist selection and handoff.

---

product-growth - Evidence-Based Growth Diagnosis

Install: claude plugin install product-growth@daymade-skills

Diagnose growth from business data and actively discovered feedback. Define the right units and cohorts, distinguish attribution from causal effects, and select improvements with measurable customer and business outcomes. Adapt to SaaS, commerce, marketplaces, developer tools and other product models.

When to use: Organic growth sources, missing channels/metrics, activation or retention problems, revenue quality, ecosystem adoption and growth-action reviews.

Use product-growth to explain our organic growth, find missing feedback channels,
and recommend the next improvement with evidence and a verification plan.

Requirements: Host file-reading and analysis tools for supplied data; permitted web/API access for live public feedback. Specialist Skills are optional adapters.

📚 Documentation: product-growth.

---

floorplan-finder - Existing Residential Floor Plans

Install: claude plugin install floorplan-finder@daymade-skills

Find original residential floor plans online and compare room connections, entrance orientation, mirrored variants, and source identity. Distinguish a user-confirmed matching layout from a documented exact-unit drawing.

When to use: Find a plan by development/building/unit, compare same-layout candidates, or recover a search stalled on inaccessible listings or ambiguous drawings.

Use floorplan-finder to find an existing plan matching this home's layout.
Keep approximate area flexible and show the original candidate images.

Requirements: Authorized web search/fetch and image-viewing tools. An interactive browser is needed for sources that require page interaction. Specialist retrieval and video Skills are optional; no bundled runtime dependencies.

📚 Documentation: floorplan-finder.

---

🎬 Interactive Demo Gallery

Want to see all demos in one place with click-to-enlarge functionality? Check out our interactive demo gallery or browse the demos directory.

🎯 Use Cases

For GitHub Workflows

Use github-ops for verified PR, issue, Actions, repository, organization-access, and API operations. Use github-review-pr when a maintainer needs a current-base code review, ownership decision, or review-gated repair/landing for one contributor PR or the complete open PR queue. An author can explicitly request the same independent read-only lens for their own contribution; implementation and authorized fixes remain with github-contributor.

For Documentation

Combine doc-to-markdown for document conversion and mermaid-tools for diagram generation to create comprehensive documentation. Use llm-icon-finder to add brand icons.

For Product Growth

Use product-growth to connect real feedback and business data to growth mechanisms, comparable metrics and verifiable actions. Use app-evolver for authorized application implementation.

For Research & Analysis

Use floorplan-finder to retrieve residential plan originals and verify them against supplied layout constraints.

Use deep-research to produce format-controlled research reports with evidence tables and citations. Combine with fact-checker to validate claims or with twitter-reader for social-source collection.

For Technology Decisions

Use tech-selection when choosing between technologies — library, framework, storage, data format, model, build-vs-buy, or architecture. It runs a gated checklist that filters candidates by business-result anchoring, prior-art inventory, and probe-based evidence, then returns candidates + trade-offs + a recommendation when multiple survive. Combine with deep-research when the landscape itself needs research before a decision can be made.

For Competitive Intelligence

Use competitors-analysis to discover, persist, update, and analyze competitor repositories with evidence-based source citations. Combine with deep-research when the landscape also needs broader market, pricing, or narrative research.

For PDF & Printable Documents

Use pdf-creator to convert markdown to print-ready PDFs with proper Chinese font support for formal documents and reports.

For Team Communication

Use teams-channel-post-writer to share knowledge and statusline-generator to track costs while working.

For Local Agent Coordination

Use the current host's native communication tools for reachable agents and sessions. Use peer-message only for uncovered local targets, cross-product/script transport, or coordination evidence that needs verification. Follow its routing contract; a denied or Held message is not a reason to switch transports.

For Repository Management & Security

Use repomix-unmixer to extract and validate repomix-packed skills or repositories. Use repomix-safe-mixer to package codebases securely, automatically detecting and blocking hardcoded credentials before distribution.

For Skill Development

Use skill-creator (see Essential Skill section above) to build, validate, and package your own Claude Code skills following best practices. Use skill-evolver for a research-driven improvement round on a specified existing Skill, from relevant evidence and trade-offs to authorized changes and result verification.

For Video Quality Analysis

Use video-comparer to analyze compression results, evaluate codec performance, and generate interactive comparison reports. Combine with youtube-downloader to compare different quality downloads.

For Media & Content Download

Use youtube-downloader to download YouTube videos and extract audio from videos with automatic workarounds for common download issues.

For Transcription & ASR Correction

Use transcript-fixer to correct speech-to-text errors in meeting notes, lectures, and interviews through dictionary-based rules and AI-powered corrections with automatic learning.

For Financial Data & Investment Research

Use the daymade-financial suite for the complete investment-research data pipeline. bigdata-skill pulls structured financials and sentiment from Bigdata.com (RavenPack) via the official SDK. financial-data-collector gathers US equity fundamentals from yfinance with validation and NO FALLBACK discipline. gangtise-copilot installs and orchestrates the full Gangtise (岗底斯) OpenAPI research suite. ashare-news-fetcher aggregates A-share news, policy, and sentiment from Chinese public sources. pharma-daily-report generates pharmaceutical sector daily reports with real-time quotes and fund-flow analysis.

For Excel & Financial Modeling Automation

Use excel-automation to create formatted workbooks, parse complex .xlsm models, and automate Excel window controls for repetitive analyst workflows.

For Visual Capture Automation on macOS

Use capture-screen to script repeatable app-window screenshots. Combine with excel-automation to generate report-ready workbook visuals.

For Meeting Documentation

Use meeting-minutes-taker to transform raw meeting transcripts into structured, evidence-based minutes. Combine with transcript-fixer to clean up ASR errors before generating minutes. Features multi-pass generation with UNION merge to avoid content loss.

For QA Testing & Quality Assurance

Use qa-expert for QA setup, test cases, execution tracking, reporting and third-party handoffs.

For Prompt Engineering & Requirements Engineering

Use prompt-optimizer to transform vague feature requests into precise EARS specifications with domain theory grounding. Perfect for product requirements documents, AI-assisted coding, and learning prompt engineering best practices. Combine with skill-creator to create well-structured skill prompts.

For Session History & File Recovery

Use read-claude-code-history to recover deleted files from previous Claude Code sessions, search for specific implementations across conversation history, or track file evolution over time. Essential for recovering accidentally deleted code or finding that feature implementation you remember but can't locate.

For Reusing Existing Work Before Producing

Use prior-work-retrieval only when the user explicitly asks to find, reuse, or reconcile earlier work. A mention of current tests, README, files, implementation, behavior, or validation is not enough. For a real prior-work request, it searches the explicit local source manifest across current code, project decisions, Skills/SOPs, meetings, archived WeChat, and conversation history; then requires source verification plus an auditable reuse/adapt/reject receipt. A zero-hit ranked search never becomes an absence claim.

For Codex Workstation Setup

Use codex-1m-context-window-setup when Codex CLI or Desktop shows about 258K context, compacts too early, or a classroom/workstation needs a reproducible long-context policy. It reads the selected model's live catalog entry, requests up to 1M raw tokens, and verifies the exact base-config values without touching the model, sandbox, approvals, plugins, or other user settings.

For Resuming Interrupted Claude Sessions

Use continue-claude-code-work to recover the last actionable request from local ~/.claude artifacts and continue implementation without reopening the original session. Combine with read-claude-code-history when you need broader cross-session search, statistics, or deleted-file recovery.

For a prior Codex CLI run, use continue-codex-work instead; it reconstructs a briefing from local Codex rollout files without replaying the full session.

For Local Conversation Evidence

Use read-claude-code-history for Claude Code inventory, exact timelines, queued human input, full-event search, and deleted-file recovery. Use read-codex-history for Codex inventory, exact prompt-ledger inputs, verified rollout identity, fork/compaction lineage, and Codex-only search. The two readers remain read-only; move to the matching continue-claude-code-work or continue-codex-work only when the user asks to execute the unfinished task.

For Web Extraction & WeChat Articles

Use scrapling-skill to install and validate Scrapling CLI, choose between static and browser-backed fetching, and extract clean Markdown from sites like mp.weixin.qq.com. Combine with deep-research to turn extracted sources into structured reports or with docs-cleaner to normalize captured article content.

For Documentation Maintenance

Use docs-cleaner to consolidate redundant documentation while preserving valuable content. Perfect for cleaning up documentation sprawl after rapid development phases or merging overlapping docs into authoritative sources.

For CLAUDE.md Optimization

Use claude-md-progressive-disclosurer to reduce CLAUDE.md bloat by moving detailed sections into references while keeping core rules visible.

For Skill Discovery & Management

Use skills-search to find and reuse Skills in configured local repositories before expanding to CCPM. Registry discovery, installation and management remain available when needed.

For LLM Evaluation & Model Comparison

Use promptfoo-evaluation to set up prompt tests, compare model outputs, and run automated evaluations with custom assertions. Use llm-eval-harness to benchmark an endpoint across speed (thinking-aware tok/s), concurrency/stability, Anthropic protocol compliance, and quality regression against your own use cases — verifying a vendor's tokens-per-second claim or vetting a newly released model before adopting it. The two compose: promptfoo for fast per-case rubric gating, llm-eval-harness for blind-judge precision and raw speed/concurrency probing.

For iOS App Development

Use developing-ios-apps to configure XcodeGen projects, resolve SPM dependency issues, and troubleshoot code signing or device deployment.

For macOS System Maintenance & Disk Space Recovery

macOS Cleaner.

For Twitter/X Content Research

Use twitter-reader to fetch tweet content without JavaScript rendering or authentication. Perfect for documenting social media discussions, archiving threads, analyzing tweet content, or gathering reference material from Twitter/X. Combine with doc-to-markdown to convert fetched content into other formats, or with repomix-safe-mixer to package research collections securely.

For Skill Quality & Open-Source Contributions

Use skill-reviewer to validate your own skills against best practices before publishing, or to review and improve others' skill repositories. Combine with github-contributor to find high-impact open-source projects, create professional PRs, and build your contributor reputation. Perfect for developers who want to contribute to the Claude Code ecosystem or any GitHub project systematically.

For Internationalization & Localization

Use i18n-expert to set up complete i18n infrastructure for React/Next.js/Vue applications, audit existing implementations for missing translation keys, and ensure locale parity between en-US and zh-CN. Perfect for teams launching products to global markets, maintaining multi-language UIs, or replacing hard-coded strings with proper i18n keys. Combine with skill-creator to create locale-aware skills, or with docs-cleaner to consolidate documentation across multiple languages.

For Network & VPN Troubleshooting

Use tunnel-doctor to diagnose and fix conflicts between Tailscale and proxy/VPN tools on macOS across multiple independent layers (route hijacking, HTTP env vars, system proxy, SSH ProxyCommand, VM/container proxy propagation, DNS resolver stall). Essential when Tailscale ping works but TCP connections fail, when git push fails with "failed to begin relaying via HTTP", or when setting up Tailscale SSH to WSL instances alongside Shadowrocket, Clash, or Surge. Also covers TUN measurement contamination — why raw probes (nc -z showing 0.00s, ping, a foreign ip-api lookup) lie while a global proxy is up, and what to trust instead.

For Product Audits

Use product-analysis for launch, UX, API or architecture audits and code-backed competitor comparisons.

For Remote Desktop & VDI Optimization

Use windows-remote-desktop-connection-doctor to diagnose Azure Virtual Desktop / W365 connection quality issues on macOS. Essential when transport shows WebSocket instead of UDP Shortpath, when RTT is unexpectedly high, or when RDP Shortpath fails after changing network locations. Combines network evidence gathering with Windows App log analysis for systematic root cause identification.

For Plugin & Skill Troubleshooting

Use claude-skills-troubleshooting to diagnose and resolve Claude Code plugin and skill configuration issues. Debug why plugins appear installed but don't show in available skills, understand the installed_plugins.json vs settings.json enabledPlugins architecture, and batch-enable missing plugins from a marketplace. Essential for marketplace maintainers debugging installation issues, developers troubleshooting skill activation, or anyone confused by the GitHub #17832 auto-enable bug.

For Tencent IMA Knowledge Base Workflows

Use ima-copilot to install the official Tencent IMA skill across Claude Code / Codex / OpenClaw, configure API credentials, detect and repair known upstream issues, and run personalized fan-out searches across all your IMA knowledge bases with priority-based boosting. The wrapper architecture means upstream upgrades never collide with your fixes — every repair is a runtime instruction, not a shipped patch. Perfect for IMA power users who switch between multiple coding agents, or for anyone who has hit the "Skipped loading skill(s) due to invalid SKILL.md" warning.

For Post-Processing Claude Code Exports

Use claude-export-txt-better to clean up /export output before archiving or sharing. The default export format hard-wraps tables, paths, and tool-call blocks at fixed column widths, which breaks readability in any viewer wider than 80 columns. The skill reconstructs the original structure and validates the fix with 53 automated checks so regressions are caught immediately.

For Personal Data Backup (Douban)

Use douban-skill to back up your Douban 书影音 (book/movie/music/game) history to CSV. Douban has no official export — the public API was shut down in 2018 and all web scraping is blocked by PoW challenges. This skill uses the same Frodo API as the official Android app, so it just works without any login or cookies. Ships with a full failure log of 7 rejected scraping approaches, saving you hours of wasted effort.

For Terraform & IaC Troubleshooting

Use terraform-skill when your terraform apply fails at a provisioner step, when fresh instances hit "docker: not found", or when multi-environment setups accidentally share snapshots. Every pattern in the skill is an exact error → root cause → copy-paste fix triple drawn from real incidents. Perfect for anyone who has lost a weekend to timing races in cloud-init, rsync connection drops in local-exec, or hardcoded domains in Caddyfiles.

For Network, Streaming & Protocol-Layer Debugging

Use debugging-network-issues when symptoms do not match the obvious cause: HTTP/2 RST_STREAM, SSE stalls at exactly 60s/100s/130s, "works sometimes but not always" failures, client-side proxy/VPN/TUN misrouting with ERR_CONNECTION_CLOSED or SSL_ERROR_SYSCALL, or anything that looks like an idle-timeout or rule-override incident through CDN / proxy / CGNAT / TUN chains. The skill replaces assumption-stacking with layered isolation experiments — running the same logical request through three or more paths that differ by one hop — plus a counter-review pattern for shipping fixes only after the hypothesis has been falsified, not just confirmed. The cognitive-trap catalog includes reverse-path / directional asymmetry and proxy-node DNS ≠ client DNS.

For Chinese TTS (StepFun StepAudio 2.5)

Use stepfun-tts for Chinese / Japanese voice synthesis with emotional control via instruction + inline () prosody. Captures the two breaking changes that ambush new StepAudio 2.5 users: voice_label removal and stricter 2.5-era censorship rules. Pair with step-tts-2 as a per-line fallback for content that triggers censorship.

For Long-Audio Transcription (StepFun StepAudio 2.5)

Use stepfun-asr for transcribing up to 30-minute Chinese / English audio in a single SSE call (32K context, ~85-101× RTF, no client-side chunking). Hides the #1 trap — the model does NOT live on /v1/audio/transcriptions; the wrong endpoint returns a misleading "model not supported" error. Combine with transcript-fixer for ASR error correction or with meeting-minutes-taker to turn long recordings into structured minutes.

For Meme & GIF Creation

Use meme-creator to put logos, avatars, or stickers onto moving objects in a video clip with frame-accurate tracking, and export the result as MP4 plus a size-budgeted GIF. Combine with youtube-downloader (or yt-dlp directly) when the source footage is still online.

📚 Documentation

Each skill includes:

• SKILL.md: Core instructions and workflows
• scripts/: Executable utilities (Python/Bash)
• references/: Detailed documentation
• assets/: Templates and resources (where applicable)

Quick Links

• floorplan-finder: Plan search and verification

• github-ops: See github-ops/references/api_reference.md for API documentation

• github-review-pr: See github-review-pr/SKILL.md for the current-base review workflow and github-review-pr/references/personal_maintainer_context.md for the explicit personal policy profile

• doc-to-markdown: See daymade-docs/doc-to-markdown/references/conversion-examples.md for conversion scenarios

• mermaid-tools: See daymade-docs/mermaid-tools/references/setup_and_troubleshooting.md for setup guide

• statusline-generator: See daymade-claude-code/statusline-generator/references/color_codes.md for customization

• teams-channel-post-writer: See teams-channel-post-writer/references/writing-guidelines.md for quality standards

• peer-message: Execution entry

• repomix-unmixer: See repomix-unmixer/references/repomix-format.md for format specifications

• skill-creator: See daymade-skill/skill-creator/SKILL.md for complete skill creation workflow

• skill-evolver: See skill-evolver/SKILL.md for target-specific research, implementation and result verification

• llm-icon-finder: See llm-icon-finder/references/icons-list.md for available icons

• cli-demo-generator: See cli-demo-generator/references/vhs_syntax.md for VHS syntax and cli-demo-generator/references/best_practices.md for demo guidelines

• cloudflare-troubleshooting: See cloudflare-troubleshooting/references/api_overview.md for API documentation

• ui-designer: See ui-designer/SKILL.md for design system extraction workflow

• youtube-downloader: See youtube-downloader/SKILL.md for usage examples and troubleshooting

• repomix-safe-mixer: See repomix-safe-mixer/references/common_secrets.md for detected credential patterns

• video-comparer: See video-comparer/references/video_metrics.md for quality metrics interpretation and video-comparer/references/configuration.md for customization options

• transcript-fixer: See daymade-audio/transcript-fixer/references/workflow_guide.md for step-by-step workflows and daymade-audio/transcript-fixer/references/team_collaboration.md for collaboration patterns

• qa-expert: See qa-expert for autonomous execution and testing standards

• prompt-optimizer: See prompt-optimizer/references/ears_syntax.md for EARS transformation patterns, prompt-optimizer/references/domain_theories.md for theory catalog, and prompt-optimizer/references/examples.md for complete transformations

• read-codex-history: See daymade-claude-code/read-codex-history/references/storage_and_portability.md for local-store selection, cross-platform paths, privacy boundaries, and diagnostics

• read-claude-code-history: See daymade-claude-code/read-claude-code-history/references/session_file_format.md for JSONL structure and daymade-claude-code/read-claude-code-history/references/workflow_examples.md for recovery workflows

• prior-work-retrieval: See daymade-claude-code/prior-work-retrieval/SKILL.md for the retrieval/verification workflow and daymade-claude-code/prior-work-retrieval/references/source-manifest.md for the explicit carrier contract

• codex-1m-context-window-setup: See daymade-codex/codex-1m-context-window-setup/SKILL.md for the doctor/apply/verify workflow and daymade-codex/codex-1m-context-window-setup/references/context_window_contract.md for model-cap, usable-window, and compaction semantics

• docs-cleaner: See daymade-docs/docs-cleaner/SKILL.md for consolidation workflows

• deep-research: See deep-research/references/research_report_template.md for report structure and deep-research/references/source_quality_rubric.md for source triage

• tech-selection: See daymade-claude-code/tech-selection/references/decision-axes.md for the 13 core filter criteria, daymade-claude-code/tech-selection/references/scoped-criteria.md for scoped supplementary criteria, daymade-claude-code/tech-selection/references/rejection-modes.md for rejection patterns and anti-patterns, and daymade-claude-code/tech-selection/references/delegation-contract.md for domain ownership and the autonomy threshold

• pdf-creator: See daymade-docs/pdf-creator/SKILL.md for PDF conversion and font setup

• claude-md-progressive-disclosurer: See daymade-claude-code/claude-md-progressive-disclosurer/SKILL.md for CLAUDE.md optimization workflow

• skills-search: See daymade-skill/skills-search/SKILL.md for CCPM CLI commands and registry operations

• promptfoo-evaluation: See promptfoo-evaluation/references/promptfoo_api.md for evaluation patterns

• developing-ios-apps: See daymade-macos/developing-ios-apps/references/xcodegen-full.md for XcodeGen options and project.yml details

• twitter-reader: See twitter-reader/SKILL.md for API key setup and URL format support

• macos-cleaner

• skill-reviewer: See daymade-skill/skill-reviewer/references/evaluation_checklist.md for complete evaluation criteria and daymade-skill/skill-reviewer/references/pr_template.md for PR templates

• github-contributor: See github-contributor/references/pr_checklist.md for PR quality checklist, github-contributor/references/project_evaluation.md for project evaluation criteria, and github-contributor/references/communication_templates.md for issue/PR templates

• i18n-expert: See i18n-expert/SKILL.md for complete i18n setup workflow, key architecture guidance, and audit procedures

• claude-skills-troubleshooting: See daymade-claude-code/claude-skills-troubleshooting/SKILL.md for plugin troubleshooting workflow and architecture

• fact-checker: See fact-checker/SKILL.md for fact-checking workflow and claim verification process

• competitors-analysis: See competitors-analysis/SKILL.md for the discover/ingest/profile/landscape workflow and competitors-analysis/references/profile_template.md for the competitor profile template

• windows-remote-desktop-connection-doctor: See windows-remote-desktop-connection-doctor/references/windows_app_log_analysis.md for log parsing patterns and windows-remote-desktop-connection-doctor/references/avd_transport_protocols.md for transport protocol details

• product-analysis: See the workflow and evidence synthesis for source verification and recommendation decisions

• excel-automation: See daymade-docs/excel-automation/SKILL.md for create/parse/control workflows and daymade-docs/excel-automation/references/formatting-reference.md for formatting standards

• capture-screen: See daymade-macos/capture-screen/SKILL.md for CGWindowID-based screenshot workflows on macOS

• continue-claude-code-work: See daymade-claude-code/continue-claude-code-work/SKILL.md for local artifact recovery, drift checks, and resume workflow

• continue-codex-work: See daymade-claude-code/continue-codex-work/SKILL.md for Codex rollout discovery, end-reason diagnosis, and continuation workflow

• scrapling-skill: See scrapling-skill/SKILL.md for the CLI workflow and scrapling-skill/references/troubleshooting.md for verified Scrapling failure modes

• ima-copilot: See ima-copilot/SKILL.md for the wrapper architecture and routing, ima-copilot/references/installation_flow.md for the install deep dive, ima-copilot/references/known_issues.md for the issue registry and repair commands, and ima-copilot/references/search_best_practices.md for the fan-out strategy and 100-result truncation details

• claude-export-txt-better: See daymade-claude-code/claude-export-txt-better/SKILL.md for the workflow, daymade-claude-code/claude-export-txt-better/scripts/fix-claude-export.py for the reconstruction algorithm, and daymade-claude-code/claude-export-txt-better/evals/ for real regression fixtures

• douban-skill: See douban-skill/SKILL.md for the export workflow and douban-skill/references/troubleshooting.md for the complete log of 7 tested scraping approaches and why each failed

• terraform-skill: See terraform-skill/SKILL.md for the full catalogue of operational traps organised by exact error → root cause → copy-paste fix

• debugging-network-issues: See debugging-network-issues/SKILL.md for the falsification-first workflow, debugging-network-issues/references/layered-isolation-experiment.md for the multi-hop isolation pattern, debugging-network-issues/references/case-sse-rst-130s.md for the SSE production case study, and debugging-network-issues/references/case-proxy-tun-cname-override.md for the client-side proxy/TUN CNAME-rule-override case study

• stepfun-tts: See stepfun-tts/SKILL.md for the Contextual TTS decision tree and stepfun-tts/references/migration_from_v2.md for the voice_label → instruction migration playbook plus the censorship rewrite list

• stepfun-asr: See stepfun-asr/SKILL.md for the SSE-endpoint workflow and the four ASR-side traps (wrong endpoint, Plan-vs-Normal key, repetition hallucination, SSE error event). stepfun-asr/references/api_reference.md documents the exact JSON request body and SSE event contract for raw HTTP integration

• llm-eval-harness: See llm-eval-harness/references/evaluation_disciplines.md for the reasoning behind each discipline (env-var keys, thinking-aware throughput, proxy isolation, probabilistic protocol verdicts) and llm-eval-harness/references/quality_blind_judge.md for the independent blind-judge quality method

• meme-creator: See meme-creator/SKILL.md for the pipeline and meme-creator/references/tracking-playbook.md for the CSRT failure taxonomy and manual-keyframe fallback

🛠 Requirements

• floorplan-finder: Authorized web retrieval and image viewing; interactive browser for sources requiring it. No bundled runtime dependencies.

• product-growth: Host file-reading/analysis tools; authorized web/API access for live feedback. No bundled runtime dependencies.

• Claude Code 2.0.13 or higher

• Codex CLI with doctor --json and debug models + uv/Python 3.11+ (for codex-1m-context-window-setup)

• gh CLI (for github-ops and github-review-pr)

• git with merge-tree --write-tree + jq (for github-review-pr)

• markitdown (for doc-to-markdown)

• mermaid-cli (for mermaid-tools)

• yt-dlp (for youtube-downloader): brew install yt-dlp or pip install yt-dlp

• FFmpeg/FFprobe (for video-comparer): brew install ffmpeg, apt install ffmpeg, or winget install ffmpeg

• pandoc + weasyprint (for pdf-creator): brew install pandoc + pip install weasyprint (or use Chrome as backend)

• VHS (for cli-demo-generator): brew install vhs

• Jina.ai API key (for twitter-reader): Free tier available at https://jina.ai/

• asciinema (optional, for cli-demo-generator interactive recording)

• ccusage (optional, for statusline cost tracking)

• Mole (optional, for macos-cleaner visual cleanup): Download from https://github.com/tw93/Mole

• repomix (for repomix-safe-mixer): npm install -g repomix

• CCPM CLI (for skills-search): npm install -g @daymade/ccpm

• Promptfoo (for promptfoo-evaluation): npx promptfoo@latest

• macOS + Xcode, XcodeGen (for developing-ios-apps)

• Codex CLI (optional, for product-analysis multi-model mode)

• uv + openpyxl (for excel-automation): uv run --with openpyxl ...

• Bigdata.com API key (for daymade-financial:bigdata-skill): bd_v2_ key from https://www.bigdata.com/

• Gangtise credentials (for daymade-financial:gangtise-copilot): accessKey + secretAccessKey from https://open.gangtise.com/

• macOS (for capture-screen and excel-automation AppleScript control workflows)

• Python 3.10+ (for the four local history read/continue skills): bundled standard-library readers and validators

• uv + Scrapling CLI (for scrapling-skill): uv tool install 'scrapling[shell]' and scrapling install for browser-backed fetches

• Node.js 18+ + curl + unzip (for ima-copilot): npx skills is fetched on demand from the npm registry; IMA OpenAPI credentials from https://ima.qq.com/agent-interface

• StepFun API key (for stepfun-tts and stepfun-asr — must be "Normal" tier, Plan keys silently fail on audio endpoints): Available at https://platform.stepfun.com/ → API Keys

• uv + an endpoint API key (for llm-eval-harness): openai and aiohttp are auto-installed via uv run --with; the key is passed by env-var name only

• FFmpeg + uv (for meme-creator): brew install ffmpeg; bundled scripts resolve their own Python deps via uv run; yt-dlp only when the source is a URL

❓ FAQ

How do I know which skills to install?

Start with skill-creator if you want to create your own skills. Otherwise, browse the Other Available Skills section and install what matches your workflow.

Can I use these skills without Claude Code?

Most marketplace skills target Claude Code. Skills in the daymade-codex suite that explicitly document Codex support can also be installed through Codex's plugin marketplace; check each Skill's requirements before use.

How do I update skills?

Follow plugin updates to refresh marketplace metadata and update the installed plugin at its existing scope. For skills moved into a suite, follow suite migration to install and verify the replacement before removing the old plugin.

Can I contribute my own skill?

Follow the contribution policy for new-skill suggestions and submissions, and the bug-fix route for existing skills.

Are these skills safe to use?

Open source makes inspection possible; it does not establish safety. Use skill-reviewer and the security checklist to inspect the relevant skill.

How do Chinese users handle API access?

We recommend using CC-Switch to manage API provider configurations. See the Chinese User Guide section above.

What's the difference between skill-creator and other skills?

skill-creator is a meta-skill - it helps you create other skills. The other skills are end-user skills that provide specific functionalities (GitHub ops, document conversion, etc.). If you want to extend Claude Code with your own workflows, start with skill-creator.

---

🤝 Contributing

This is a curated marketplace of our own skills — see CONTRIBUTING.md for the full policy. In short:

1. Bug reports and bug-fix PRs are welcome — open an issue first, then a scoped fix
2. New-skill PRs are not accepted — publish your own marketplace instead (the format is open)
3. Feedback on skill quality is always appreciated

Skill Quality Standards

All skills in this marketplace follow:

• Imperative/infinitive writing style
• Progressive disclosure pattern
• Proper resource organization
• Comprehensive documentation
• Tested and validated

📄 License

This marketplace is licensed under the MIT License - see the LICENSE file for details.

⭐ Support

If you find these skills useful, please:

• ⭐ Star this repository
• 🐛 Report issues
• 💡 Suggest improvements
• 📢 Share with your team

🔗 Related Resources

• Claude Code Documentation
• Agent Skills Guide
• Plugin Marketplaces
• Anthropic Skills Repository

📞 Contact

• GitHub: @daymade
• Email: daymadev89@gmail.com
• Repository: daymade/claude-code-skills

---

Built with ❤️ using the skill-creator skill for Claude Code