Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 80 149 files from 1 774 authors, of which 62 489 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Install skills from github.com/oaustegard/claude-skills into /mnt/skills/user. Use when user mentions "install skills", "load skills", "add skills", "update skills", "refresh skills", or references a skill not currently installed.
Install and drive Google's Antigravity CLI (`agy`) as a non-interactive sub-agent. Use when orchestrating agy, running Antigravity agents from a script or sandbox, delegating a task to Google's agent harness, or wanting a Gemini-backed peer agent alongside Claude.
Discovers and indexes Python code in skills, enabling cross-skill imports. Use when importing functions from other skills or analyzing skill codebases.
Multi-conversation methodology for iterative stateful work with context accumulation. Use when users request work that spans multiple sessions (research, debugging, refactoring, feature development), need to build on past progress, explicitly mention iterative work, work logs, project knowledge, or cross-conversation learning.
Generate guardrailed UI from natural language. Emits constrained JSON that a Preact runtime renders. Use when the request is for a dashboard with metrics, charts, or tables; an admin panel; a data visualization interface; or a form-based application.
Execute programs on a compiled transformer stack machine where every instruction fetch and memory read is a parabolic attention head. Demonstrates that transformer attention + FF layers can implement a working computer. Use when user mentions "llm-as-computer", "lac", "stack machine", "compiled transformer", "percepta", "parabolic attention", "execute program", or asks to run/trace programs on the transformer executor.
Generates WAFFLES Declarations for social media posts — preemptive lists of what a post does NOT say. Use when users mention WAFFLES, ask for clarifications on their post, want to prevent misinterpretation, or request disclaimers for controversial/nuanced takes.
Generate navigable semantic maps from PDF documents. Extracts section structure via font analysis, then runs LLM extraction per section for claims, symbols, and dependencies — all page-anchored. Produces _MAP.md (progressive disclosure), .symbols.json (definition index), .anchors.json (claim references), and a _USAGE.md snippet for CLAUDE.md. Use when analyzing papers, specs, or legal docs; when asked to "map this document", "index this PDF", "what does this paper say"; or when a coding agent needs grounded reference material from a PDF source. Analogous to tree-sitting but for prose documents.
Disciplined, validation-gated revision of an EXISTING skill so each edit is a measured improvement rather than a guess. Use when editing, revising, or tuning a skill that already exists and there is evidence it underperforms (observed failures, drift, complaints) — invoke by name, or have versioning-skills / creating-skill defer to it before applying edits. Not for authoring a brand-new skill from scratch (use creating-skill) or one-off prose.
Open a GitHub PR via API as a flowing graph — branch + push + create_pr + mergeable poll, with structural protection against pushing to main/master/etc. Use when a Claude Code or Claude.ai container needs to land changes on GitHub via the API (no git CLI required) and the prose "create branch, push, open PR, wait for mergeable" workflow keeps drifting under context pressure.
DEPRECATED — superseded by tree-sitting. This skill generated persistent _MAP.md files; tree-sitting does the same AST extraction at runtime (~700ms for 250 files) with no artifact to write, commit, or keep in sync. Use tree-sitting for "map this codebase", "explore repo", "understand structure", or any code navigation task. The final working version is archived at release mapping-codebases-v0.8.0.
Generate behavioral/feature documentation for web apps using code-first analysis. Reads source code via tree-sitting to produce _FEATURES.md, with optional visual verification via browser automation. Companion to tree-sitting. Use when documenting app behavior, creating feature inventories, generating behavioral ground truth for agents, or before modifying UI code. Triggers on "map features", "document app behavior", "feature inventory", "what does this app do".
Orchestrates parallel API instances, delegated sub-tasks, and multi-agent workflows with streaming and tool-enabled delegation patterns. Routes by surface — native subagents in Cowork and Claude Code, httpx fan-out on claude.ai — and covers Gemini delegation via the Cloudflare AI Gateway on every surface. Use for parallel analysis, multi-perspective reviews, or complex task decomposition.
>- Skill-aware orchestration with context routing. Decomposes complex tasks into skill-typed subtasks, extracts targeted context subsets, executes subagents in parallel, and synthesizes results. Self-answers trivial lookups inline. No SDK dependency — uses raw HTTP via httpx. Use when tasks require multiple analytical perspectives, when context is large and subtasks only need portions, or when orchestrating-agents spawns too many redundant subagents.
>- Interactive codebase orientation for human learning. Companion to exploring-codebases (which builds Claude's understanding); this skill builds the user's understanding through guided exercises grounded in learning science. Uses the same tree-sitting + featuring pipeline but synthesizes into interactive teaching via HTML artifacts rather than analysis documents. Triggers on "orient me to this repo", "teach me this codebase", "help me understand this code", "learning orientation", or when the user wants to build genuine comprehension of an unfamiliar codebase rather than just getting work done in it.
Interpret video content visually by sampling frames into timestamped contact sheets that can be read as images. Use when: user asks what happens in a video; asks to summarize, describe, review, or QA video content or footage; asks about scenes, actions, people, or objects in a video; needs a storyboard-style overview of a clip; asks to find where something occurs in a video. Triggers on 'watch this video', 'what's in this video', 'summarize the video', 'describe the footage', 'contact sheet', 'storyboard', 'review this clip', 'find the scene where', 'scene detection', 'shot boundaries', 'detect cuts', 'dwell points'. For converting, trimming, or transcoding video, use processing-video instead.
Exhaustive problem space exploration using the MIT Synthetic Neurobiology "tiling tree" method. Partitions a problem into MECE (Mutually Exclusive, Collectively Exhaustive) subsets recursively via parallel subagents, then evaluates leaf ideas against specified criteria. Use when users say "tiling tree", "tile the solution space", "exhaustively explore approaches to", "what are all the ways to", or request a MECE breakdown of a problem. Requires orchestrating-agents skill.
Analyze AI/ML technical content (papers, articles, blog posts) and extract actionable insights filtered through enterprise AI engineering lens. Use when user provides URL/document for AI/ML content analysis, asks to "review this paper", or mentions technical content in domains like RAG, embeddings, fine-tuning, prompt engineering, LLM deployment.
Check that a document's claims about code are actually true by reading the prose, the code, and the tests and reporting (or fixing) where they disagree. Use whenever the user wants to verify a README, guide, spec, or docstring still matches the code; whenever they mention documentation drift, doc-code sync, "is this still accurate", stale docs, or keeping docs/tests/code consistent; before publishing or merging a docs change; or as a periodic doc-accuracy sweep. The agent reads the prose's meaning directly — there is no claim-comment DSL to maintain. Pairs with TDD — the test suite is the deterministic behavioral gate, this skill is the semantic prose-vs-reality review.
Semantic Python code queries via a stdio LSP client driving pyright-langserver. Provides binding-resolved go-to-definition, find-references, hover types, type diagnostics, file symbol outlines, and project-wide symbol search — name resolution and type inference that tree-sitter and ripgrep cannot do. Use when you need to follow an import to a definition, find all real uses of a symbol (excluding same-named-but-unrelated ones), get an inferred type, surface type errors, outline a file, or search symbols across a project. Triggers on "go to definition", "find references", "what type is", "resolve this symbol", "symbol outline", "find symbol in project", "pyright", "type-check this file".
Apply semi-formal certificate reasoning to code analysis — patch verification, fault localization, patch equivalence. Use when reviewing patches, hunting bugs across scopes, comparing fixes, or when code reasoning requires tracing execution across files/modules. Triggers on code review, bug localization, patch comparison, name shadowing, scope analysis, regression checking.
>- Binding-resolved Python symbol queries — find all true callers (--refs), go-to-definition (--def), or inferred signature (--hover) of a .py symbol via pyright, excluding same-named false positives that text grep cannot. Use when a task needs ALL callers/users of a Python symbol or its real definition and text matching would over-match. For everything else — literal tokens, regex patterns, concept/natural-language search, any repo on real issue-localization tasks, the semantic and indexed-regex tiers tied or lost against naive rg at 4-60x the wall-clock cost. Those tiers remain available below but are NOT recommended as a default.
In-process semantic search over text files or in-memory strings, using Gemini embeddings via the CF AI Gateway. Use when user wants fuzzy/conceptual search where exact-keyword grep would miss — "sessions discussing regulatory constraints", "code about retry logic", "notes mentioning burnout even if the word isn't there". Complements searching-codebases (regex/AST) and extracting-keywords (YAKE). Do NOT use when an exact string/regex match is what's wanted — grep/rg wins on speed and precision there.
AST-powered code navigation via tree-sitter. Auto-scans codebases and provides progressive-disclosure tree views with symbol search, source retrieval, and reference finding. Each invocation is self-contained — no cross-process state. Use when exploring unfamiliar repos, navigating code, or needing fast symbol lookup. Triggers on "map this codebase", "explore repo", "find symbol", "navigate code", "tree-sitter", or when starting work on an unfamiliar repository.
Query, filter, and transform Markdown structurally with mq — a jq-like CLI for Markdown. Use to extract headings/sections/code-blocks/links from .md files, build a table of contents, pull code blocks of a given language, slice or reshape LLM prompt/output Markdown, or batch-transform docs. Triggers on "extract sections from this markdown", "get all the code blocks", "jq for markdown", "mq", or any structural query over Markdown that grep/Read can't do cleanly.
File-upload bridge for Claude Code on the Web. CCotw has no native file mount; this skill creates a throwaway GitHub branch the user can drop files onto via the github.com web UI, then fetches them locally on the next turn. Use when the user wants to upload, share, or send files into the session, or when a task clearly needs files the user has on disk that aren't in the repo.
Maintain a structured task list for the current session. Use proactively when a request requires 3+ distinct steps, the user provides multiple items, or complex work benefits from explicit progress tracking. Storage persists via Muninn config across container death. Adapted from Claude Code's TodoWrite tool.
Maintains a structured running-notes document during long work sessions. Use when the user says "session notes", "update notes", "start session notes", "show session notes", or when you recognize the current session has accumulated enough state (decisions, corrections, files touched, errors) that it risks being lost under context pressure. Stores notes as a procedure memory tagged [session-memory, active] so they survive container death within the same session thread.
Systematic research methodology for building comprehensive, current knowledge on any topic. Requires web_search tool. Use when questions require thorough investigation, recent developments post-cutoff, synthesis across multiple sources, or when Claude's knowledge may be outdated or incomplete. Triggered by "Research", "Investigate", "What's current on", "Latest info on", complex queries needing validation, or technical topics with recent changes.
Image processing toolkit awareness. Use when: user uploads images for manipulation, requests format conversion, batch processing, compositing, resizing, optimization, analysis, effects, metadata inspection, montages, animated GIFs, color correction, or any image-related task. Also use when working with screenshots, photos, diagrams, icons, or visual assets. Triggers on 'resize', 'crop', 'convert', 'compress', 'optimize', 'thumbnail', 'watermark', 'montage', 'collage', 'gif', 'sprite sheet', 'color space', 'metadata', 'EXIF', 'compare images', 'diff', 'overlay', 'composite', 'batch process', 'image analysis', 'histogram', 'blur', 'sharpen', 'rotate', 'flip', 'border', 'shadow', 'round corners', 'favicon', 'icon set'.
Audio and video processing with ffmpeg. Use when: user asks to convert, trim, merge, compress, or transcode video or audio files; extract audio from video; create GIFs or animated WebP from video; add subtitles or watermarks to video; change video resolution, framerate, or codec; normalize audio loudness; extract frames from video; concatenate clips; create thumbnails from video; strip or add audio tracks; convert between audio formats (MP3, AAC, FLAC, Opus, WAV); adjust volume; apply video filters; stabilize shaky video; generate waveform or spectrum visualizations; probe media file metadata. Triggers on 'ffmpeg', 'video', 'audio', 'transcode', 'MP4', 'MKV', 'WebM', 'MP3', 'AAC', 'FLAC', 'Opus', 'WAV', 'GIF from video', 'extract audio', 'add subtitles', 'video to gif', 'compress video', 'trim video', 'merge videos', 'normalize audio', 'framerate', 'resolution', 'bitrate', 'codec', 'ffprobe', 'waveform', 'spectrogram'.
Augmented vision tools for analyzing images beyond native visual capabilities. Use when tasked with describing images in detail, reproducing images as SVGs, identifying subtle features, comparing image regions, reading degraded text, or any task requiring careful visual inspection. Also use when the image-to-svg skill needs ground truth about colors, shapes, or boundaries.
Preprocesses photographed sheets of many business cards — slicing each into overlapping high-resolution tiles and de-glaring them with container tooling (OpenCV/ImageMagick) — then reads every card via cheap parallel temperature-0 API calls (Haiku or Sonnet) using a distilled extraction prompt, and writes deduped contact fields to a CSV. Use when a user has photos or scans holding multiple business cards per image, mentions glare or unreadable cards, batch card transcription, contact extraction, or wants to read many cards without an expensive in-conversation pass. Triggers on 'business cards', 'card scan', 'extract contacts', 'read these cards', 'card glare', 'too many cards per photo'.
Portrait Mode for SVGs — foveated vectorization with 4-zone selective detail. Combines vision annotations, MediaPipe segmentation/landmarks, and optional saliency. Like phone portrait mode, but vectorized. Use when vectorizing a portrait or photo where subject detail should outrank background detail.
Reads the visual content of slides, pages, and images the way a human would, not just their embedded text. Use when a PPTX or PDF has image slides, screenshots, charts, scanned figures, or flattened-to-image layouts that the built-in pptx/pdf skills read as empty; when asked to transcribe, describe, OCR, or extract what is shown in an image, slide deck, or document page; or when embedded-text extraction returned little or nothing from a visually rich file. Triggers on 'read this deck', 'what's on these slides', 'transcribe', 'OCR', 'extract text from image', 'describe this chart/diagram', .pptx/.pdf/.png/.jpg with visual content.
REQUIRED for all skill development. Automatically version control every skill file modification for rollback/comparison. Use after init_skill.sh, after every str_replace/create_file, and before packaging.
Write effective instructions for Claude: project instructions, standalone prompts, and skill content. Use when users need help writing prompts, setting up project instructions, choosing between instruction formats, or improving how they communicate with Claude. Covers writing principles, model-aware calibration, and format selection. For building and testing complete skills, use skill-creator instead.
Sort grocery lists by aisle order using store aisle sign photos. Build aisle maps from uploaded images, match items to aisles, and output optimized shopping routes. Use when users upload aisle sign photos, request grocery list sorting, want shopping trip optimization, need store layout mapping, or mention grocery list organization.
Browser automation via webctl CLI in Claude.ai containers with authenticated proxy support. Use when users mention webctl, browser automation, Playwright browsing, web scraping, or headless Chrome in container environments.
DEPRECATED - Use browsing-bluesky skill instead. Sample and analyze Bluesky firehose to identify trending topics and content clusters. Use when user asks about "what's happening on Bluesky", "Bluesky trends", "zeitgeist", "firehose analysis", or wants to see real-time topic clusters from the network.
Design Apple-style iOS/macOS interfaces following Human Interface Guidelines. Creates HIG-compliant components with SF Symbols, San Francisco typography, and proper accessibility. Supports optional modern effects. Use when designing Apple-style UI, iOS/macOS interfaces, HIG-compliant components, or implementing design system specifications.
Embedded hardware debugging workflow for probe-rs targets using embedded-debugger-mcp. Use when Codex or Claude Code needs to inspect debug probes, validate embedded debugger setup, start the MCP server, guide a user through ARM Cortex-M/RISC-V flashing/debugging/RTT workflows, or operate without installing an MCP client by using the CLI plus prompts.
Ask clarifying questions to users via interactive TUI. Use when you need user input on preferences, implementation choices, or ambiguous instructions.
Scan current session for skills that need creating or updating, then apply changes
Save session state and generate a return prompt (quick or full mode)
Convert markdown to a beautifully styled, shareable HTML page and upload to S3-compatible storage
Turn a session into a high quality Claude skill with research and planning
Session retrospective - analyze what happened, find friction, auto-implement quick wins, publish report
Socratic teaching loop for any Claude Code session — quiz yourself on what actually happened, confirm mastery item by item, and don't finish until everything's locked in
Execute Step 5 of the GenUI Workshop, integrating the GenUI package into the Flutter app.
Creates a shell script to build and run the Flutter web app on Cloud Shell using a local HTTP server.
Execute Step 6 of the GenUI Workshop, creating the weather input widget and updating the system prompt.
Execute Step 3 of the GenUI Workshop, creating the empty Flutter project and adding dependencies.
> Orchestrates the discovery, mapping, design, and implementation of a premium Flutter-based frontend for an Agent Development Kit (ADK) agent written with Python.
Execute Step 4 of the GenUI Workshop, setting up the GenUI scaffolding and initial chat interface.
Execute Step 7 of the GenUI Workshop, creating the weather card widget and fake forecast data.
Use when writing non-subsystem tests in assisted-service.
Build and code-generate assisted-service using skipper with podman. Use when running make targets, building the service, regenerating code from swagger or CRDs, running linters, or any containerized development task. Also use when the user encounters build errors, needs to set up their development environment, or asks about how to run tests.
Finds, evaluates, fills, submits, and tracks a candidate's own job applications using a verified resume, evidence-based targeting, secure local profile storage, and browser automation. Use for onboarding or migrating a job-search profile, searching active roles, assessing a posting, applying to an authorized URL or batch, recording outcomes, or reviewing application effectiveness.
Claude Skills — questions
Answers built from the skills we actually parsed.