.agents

Guide for adding new AI provider documentation. Use when adding documentation for a new AI provider (like OpenAI, Anthropic, etc.), including usage docs, environment variables, Docker config, and image resources. Triggers on provider documentation tasks.

Analyzes, generates, and enhances CLAUDE.md files for any project type using best practices, modular architecture support, and tech stack customization. Use when setting up new projects, improving existing CLAUDE.md files, or establishing AI-assisted development standards.

Create, review, and iterate on Claude Code skills using Anthropic's official best practices and latest API documentation. Use when creating skills, reviewing existing skills, writing skill descriptions, designing skill architecture, or when user says "create a skill", "review my skill", "skill best practices", "skill description help". Do NOT use for creating plugins, commands, agents, or hooks - use plugin-dev skills for those.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

Install a pre-commit hook that scans .specstory/history for secrets before commits. Run when user says "set up secret scanning", "install specstory guard", "protect my history", or "check for secrets".

Analyze your SpecStory AI coding sessions in .specstory/history for yak shaving - when your initial goal got derailed into rabbit holes. Run when user says "analyze my yak shaving", "check for rabbit holes", "how distracted was I", or "yak shave score".

Complete project architecture and structure guide. Use when exploring the codebase, understanding project organization, finding files, or needing comprehensive architectural context. Triggers on architecture questions, directory navigation, or project overview needs.

Use when the user asks to "audit my codebase", "act as a project architect", "find open source alternatives", "enhance this project", "scale the architecture", or "generate missing documentation". Use when the user needs structural improvements, security audits, paradigm shifts, or broad ecosystem enhancements beyond simple refactoring.

Automate macOS tasks using AppleScript, shell scripts, Shortcuts, and system commands. Use when working with macOS automation, AppleScript, Finder operations, system preferences, app automation, file management, or macOS-specific workflows.

Create, debug, and optimize shell scripts (bash, zsh, sh) for automation, system administration, file processing, and CLI workflows. Use when writing shell scripts, automating tasks, processing files, system administration, or working with command-line tools.

Expert guidance for Google Ads Script development including AdsApp API, campaign management, ad groups, keywords, bidding strategies, performance reporting, budget management, automated rules, and optimization patterns. Use when automating Google Ads campaigns, managing keywords and bids, creating performance reports, implementing automated rules, optimizing ad spend, working with campaign budgets, monitoring quality scores, tracking conversions, pausing low-performing keywords, adjusting bids based on ROAS, or building Google Ads automation scripts. Covers campaign operations, keyword targeting, bid optimization, conversion tracking, error handling, and JavaScript-based automation in Google Ads editor.

Comprehensive Google Tag Manager guide covering container setup, tags, triggers, variables, data layer, debugging, custom templates, and API automation. Use when working with GTM implementation, configuration, optimisation, troubleshooting, or any GTM-related tasks.

Linear issue management guide. Use when working with Linear issues, creating issues, updating status, or adding comments. Triggers on Linear issue references (LOBE-xxx), issue tracking, or project management tasks. Requires Linear MCP tools to be available.

Format and validate markdown documents with auto-generated TOC, frontmatter, structure validation, and cross-reference linking. Export to GitHub/CommonMark/Jekyll/Hugo.

Audit and improve CLAUDE.md files in repositories. Use when user asks to check, audit, update, improve, or fix CLAUDE.md files. Scans for all CLAUDE.md files, evaluates quality against templates, outputs quality report, then makes targeted updates. Also use when the user mentions "CLAUDE.md maintenance" or "project memory optimization".

Convert audio files between formats (MP3, WAV, FLAC, OGG, M4A) with bitrate and sample rate control. Batch processing supported.

Remove backgrounds from images using segmentation. Support for color-based, edge detection, and AI-assisted removal methods. Batch processing available.

FFmpeg CLI reference for video and audio processing, format conversion, filtering, and media automation. Use when converting video formats, resizing or cropping video, trimming by time, replacing or extracting audio, mixing audio tracks, overlaying text or images, burning subtitles, creating GIFs, generating thumbnails, building slideshows, changing playback speed, encoding with H264/H265/VP9, setting CRF/bitrate, using GPU acceleration, creating storyboards, or running ffprobe. Covers filter_complex, stream selectors, -map, -c copy, seeking, scale, pad, crop, concat, drawtext, zoompan, xfade.

Generate app icons and favicons in multiple sizes from a single source image. Support for iOS, Android, web favicon, and social media formats.

Extract EXIF metadata from images including GPS coordinates, camera settings, and timestamps. Map photo locations and strip metadata for privacy.

Cut video segments by timestamp, split videos into chunks, trim start/end, and extract specific scenes with precise frame control.

Convert between data formats (JSON, CSV, XML, YAML, TOML). Handles nested structures, arrays, and preserves data types where possible.

Use when asked to compare multiple ML models, perform cross-validation, evaluate metrics, or select the best model for a classification/regression task.

Generate extractive summaries from long text documents. Control summary length, extract key sentences, and process multiple documents.

Analyze a codebase and recommend Claude Code automations (hooks, subagents, skills, plugins, MCP servers). Use when user asks for automation recommendations, wants to optimize their Claude Code setup, mentions improving Claude Code workflows, asks how to first set up Claude Code for a project, or wants to know what Claude Code features they should use.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, or extract information from web pages.

Use when asked to profile Python code performance, identify bottlenecks, measure execution time, or analyze function call statistics.

Automated code review for pull requests using specialized review patterns. Analyzes code for quality, security, performance, and best practices. Use when reviewing code changes, PRs, or doing code audits.

Connect Codex to any app. Send emails, create issues, post messages, update databases - take real actions across Gmail, Slack, GitHub, Notion, and 1000+ services.

Compare document similarity using TF-IDF, cosine similarity, and Jaccard index. Use for plagiarism detection, duplicate finding, or content matching.

Electron desktop development guide. Use when implementing desktop features, IPC handlers, controllers, preload scripts, window management, menu configuration, or Electron-specific functionality. Triggers on desktop app development, Electron IPC, or desktop local tools implementation.

This skill helps launch and configure the Chrome DevTools MCP server, giving Claude visual access to a live browser for debugging and automation. Use when the user asks to set up browser debugging, launch Chrome with DevTools, configure chrome-devtools-mcp, see what my app looks like, take screenshots of my web application, check the browser console, debug console errors, inspect network requests, analyse API responses, measure Core Web Vitals or page performance, run a Lighthouse audit, test button clicks or form submissions, automate browser interactions, fill out forms programmatically, simulate user actions, emulate mobile devices or slow networks, capture DOM snapshots, execute JavaScript in the browser, or troubleshoot Chrome DevTools MCP connection issues. Supports Windows, Linux, and WSL2 environments.

The Fifteen-Factor App methodology for modern cloud-native SaaS applications. This skill should be automatically invoked when planning SaaS tools, product software architecture, microservices design, PRPs/PRDs, or cloud-native application development. Extends the original Twelve-Factor App principles with three additional factors (API First, Telemetry, Security). Trigger keywords include "fifteen factor", "12 factor", "SaaS architecture", "cloud-native design", "application architecture", "microservices best practices", or when in a planning/architecture session.

Generate descriptive commit messages by analyzing git diffs. Use when the user asks for help writing commit messages or reviewing staged changes.

Identifies and removes AI writing patterns to make text sound natural and human-written. Use when user says "humanise this", "make this sound less AI", "this reads like a robot wrote it", "de-AI this text", "remove AI patterns", "make this more natural", "clean up this AI-generated text". Detects and fixes 28 patterns based on Wikipedia's "Signs of AI writing" guide - inflated language, promotional tone, AI vocabulary, em dash overuse, filler phrases, sycophantic tone, placeholder text, formulaic structure. Do NOT use for grammar-only proofreading, spell checking, or rewriting text that is already clearly human-written.

UI copy and microcopy guidelines. Use when writing UI text, buttons, error messages, empty states, onboarding, or any user-facing copy. Triggers on i18n translation, UI text writing, or copy improvement tasks. Supports both Chinese and English.

Plan, focus, and complete work with energy management, time blocking, goals, projects, tasks, habits, reviews, priorities, and context-specific productivity systems; use when (1) the user needs help with productivity, focus, time management, planning, priorities, goals, projects, tasks, habits, or reviews; (2) they want a reusable structure or workspace for organizing work; (3) ongoing work should be routed through a dedicated productivity framework.

Generate comprehensive Product Requirement Plans (PRPs) for feature implementation with thorough codebase analysis and external research. Use when the user requests a PRP, PRD, or detailed implementation plan for a new feature. Conducts systematic research, identifies patterns, and creates executable validation gates for one-pass implementation success. Do NOT use for client discovery, requirements gathering, or scope definition - use scope-clarifier for those.

Guide for using Recent Data (topics, resources, pages). Use when working with recently accessed items, implementing recent lists, or accessing session store recent data. Triggers on recent data usage or implementation tasks.

Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks.