Explore AI Agent Skills & Claude Prompts
Discover open-source agent skills for Claude Code, Codex, ChatGPT, and any tool that uses SKILL.md.
Enter through keywords, occupations, creators, and GitHub sources to see what kinds of skills are emerging across domains.
Use the same catalog through the API
Connect 381,784 public skills to your own search, analytics, or agent workflow with the REST API.
Querying local SQLite index...
save-session
by gurusharan3107Snapshot tactical working context to `.claude/session-data/CURRENT.md` for the next session to resume from. Triggers: `/save-session`, 'save session', 'save progress', 'checkpoint'. Captures what `docs/goal/STATUS.md` does NOT — current intent, next concrete action, open blockers, mid-session learnings, key files touched, useful commands. Writes atomically via Bash heredoc (no Read→Write context bloat). Project-local replacement for the user-global save-session, which was removed because its body triggered compaction near context limit.
self-optimize
by gurusharan3107Analyze recent Claude Code session transcripts and git history to surface recurring mistakes, map each to its root cause and the correct surface to fix it, and apply targeted edits. Use when the operator asks 'what mistakes am I making', 'what keeps going wrong', 'self-optimize', 'analyze recurring issues', 'improve yourself', 'encode learnings from sessions', 'why do I keep correcting you', 'update surfaces for recurring mistakes', 'self-introspect on what went wrong', 'surface recurring patterns', or any variant pairing session analysis / recurring mistakes / self-improvement / surface update / recurring corrections language with execution. Also use proactively at session entry after a >3-day gap when memory contains unresolved correction entries. Reads ~/.claude/projects/ transcripts via scripts/cluster.py (which consumes the bundled scripts/analyze-sessions.mjs output) and git log for fix-commit patterns. Clusters corrections into themes with frequency counts, maps each theme to: (a) root cause — missing
start
by gurusharan3107Primary session-entry skill. ONE command at the start of every new session in this repo, whether or not the prior session ran /save-session. Loads framework context (AGENTS.md root + docs/goal/README.md), STATUS Current Position + Last Update + top Recent Decisions, recent git log, optional drift warnings from check_status_drift.py, and — only if .claude/session-data/CURRENT.md exists and is <48h old — the tactical handoff from the prior session. Synthesizes ONE briefing message and waits for direction; never auto-executes. Triggers: `/start`, 'start', 'begin', 'hi', 'hi how are you', 'where are we', 'what's next', 'provide status of where we are', 'can you check AGENTS.md and docs/goal/README.md', 'can you first check the AGENTS.md and docs/goal/README.md', 'check the framework', 'what's the status', or any phrasing that pairs 'check' / 'read' / 'load' with AGENTS.md / docs/goal/README.md / STATUS.md at session entry. Sister skill to `/resume-session` — both converge to the same loaded state: `start` runs th
status
by gurusharan3107Owns docs/goal/ governance + the operator overview page. CONSULT THIS SKILL BEFORE editing ROADMAP.md or STATUS.md — it defines the maintenance contract (where pending vs completed items live, compactness budgets, what is necessary, how it synthesizes into goal-overview.html). Six lanes: contract (read the rules before editing), lint (check ROADMAP/STATUS obey them + HTML is in sync), update (deterministically regenerate the overview's live numbers + priorities), open (launch the page), build (autonomous build loop — work through ROADMAP items in order, build each, mark [x], commit, /status update, repeat until done or dashboard-gated), test (for each closed [x] item with a pending T: token, run the appropriate test, upgrade the token to passed/na, commit, then /status update). Triggers: "/status open|update|lint|build|test", "update the goal overview", "refresh goal-overview", "before I edit the ROADMAP/STATUS", "lint the goal docs", "how should I update the roadmap", "start building", "work through the road
agent-feedback-artifact
by gurusharan3107Use when the user wants in-page annotation widget on HTML artifacts, marker-local chat, or comment-triggered agent work. Add, serve, queue, and process marker feedback. Triggers: annotation, feedback, marker, artifact.
autoresearch
by gurusharan3107Single entry point for the M3.5 Track B autoresearch optimization loop. Three lanes — Baseline (establish σ-floor), Iterate (pick idea → run → verdict), Fix (source-patch a gap the loop surfaced and can't patch itself). On invocation, ALWAYS asks the operator which lane via AskUserQuestion; only skips the question when the typed prompt unambiguously names one (e.g. 'run baseline', 'iterate on idea 4', 'fix the telemetry gap'). Use whenever the operator asks to 'run autoresearch', 'start the loop', 'kick off baseline', 'run baseline', 'try the next optimize idea', 'iterate on optimization #N', 'compare a candidate', 'add a new optimize idea', 'fix the gap raised by autoresearch', 'address the autoresearch blocker', 'patch the telemetry/contract/schema gap', or any variant pairing autoresearch / loop / optimize / baseline / fixture / gap / blocker language with execution. ALSO use proactively when a `knowledge-base` refresh shifts the Claude Agent SDK rubric (re-baseline → Baseline), whenever STATUS.md says ACT
builder-test
by gurusharan3107Use when the user says "test the builder", "verify my fix", "run builder tests", "run builder test", "did my change work", "check agent behavior", "verify the builder", "test my change", "check if the fix worked", or "did the agent behave correctly after my change". Runs a 6-phase verification loop against the running builder: preconditions → static (Pyright + bad-string grep) → unit (behavioral assertions on pure functions) → integration (REST API smoke tests) → E2E (submit operator instruction, observe session behavior, verify side-effects) → verdict. Produces a PASS/WARN/FAIL table with cited evidence per phase. Optional scope: /test <phase> (e.g. "static", "unit", "integration", "e2e") to run a single phase only.
create-skill-backup
by gurusharan3107ARCHIVED snapshot of the historical create-skill (Create/Audit/Optimize lane model, from commit 54c1e0f^), kept for reference only. Do NOT activate for skill-authoring work — use the live `create-skill` skill instead. This copy exists solely as a preserved backup.
create-skill
by gurusharan3107Use when the user asks to create a new agent skill, scaffold a skill directory, add a reusable workflow as a skill, build a slash command, improve an existing skill, or "turn this into a skill". Three lanes — Create (scaffold a new skill following the webwright pattern: preflight → explore → author → validate → closeout with introspection loop), Refine (improve an existing skill's structure, description, or content), Audit (check a skill against agentskills.io spec and project standards). Output skills are ready to activate, self-verify, and self-improve — not blank stubs. Also triggers on: "make a skill for X", "create a slash command for Y", "refine the Z skill", "why isn't my skill activating", "add a skill that does X".
elon
by gurusharan3107Evaluate a codebase or diff through the Musk engineering algorithm — question requirements, DELETE, simplify, accelerate cycle time, automate last — scored against hard git/grep evidence, not impressions. Produces an evidence-grounded 6-criterion scorecard (deletion ratio, idiot index, requirements sanity, cycle time, automate-last, vertical ownership) plus a VERIFIED, import-checked, tiered deletion list with per-cut coupled-edit checklists. Orchestrates the built-in /code-review command (find what's broken) and /simplify command + code-simplifier agent (reduce mass, apply cuts) as phases inside the algorithm. Optionally renders the report via html-artifact. Use when the operator says 'put the Elon hat on', 'elon', 'musk hat', 'first-principles audit', 'what can we delete', 'deletion ratio', 'idiot index', 'is this codebase too big', 'what should we delete', 'audit this codebase for bloat', 'find dead/duplicate code to cut', or pairs delete/cut/prune/simplify with first-principles/Musk/should-this-exist lang
hermes-chrome-bridge
by gurusharan3107Install the Hermes Chrome Bridge — deploys the hermes_chrome plugin (extension + native host) from the repo, wires native messaging for the current platform (macOS or WSL2), and installs the hermes-chrome operate skill globally. Run once per machine. Triggers: "install hermes chrome", "set up the chrome bridge", "set up hermes-chrome", "wire native messaging", "bootstrap the browser bridge".
hermes-chrome
by gurusharan3107Use this skill to operate Chrome through the Hermes bridge — navigate pages, click with a visible animated cursor, take screenshots, zoom into page regions, read page content, fill forms, and interact with authenticated browser state. Use even when the user doesn't say "Chrome" or "browser" explicitly: applies any time they want to verify, test, screenshot, click, or interact with a running web app or authenticated site. Also use when the bridge is misbehaving, an action is failing, the cursor is not showing, the snapshot is noisy, screenshots time out, or any Hermes Chrome behaviour needs fixing. Triggers: "use Chrome", "open Chrome", "take a screenshot", "click X in the browser", "navigate to", "fill out a form", "verify in Chrome", "test in browser", "Chrome with my login", "authenticated browser", "hermes chrome", "browser testing", "bridge not working", "fix the extension", "cursor not showing", "optimize the bridge". Requires hermes-chrome-bridge installed.
Browse Agent Skills by Occupation
23 major groups · 867 SOC occupations
Browse by Category
Explore agent skills organized by their primary use case
Explore the agent skills ecosystem by occupation and creator
SkillMD is not just a keyword search box. It is an open map that organizes public skills by occupation, creator, and repository, helping you see which workflows, judgment criteria, and domain habits people are writing for AI agents.
Then follow creators and GitHub repositories back to the source: compare the skills a team maintains, whether the repo is active, and how the README frames the work before you open, install, or reuse anything.
Use it three ways: learn an unfamiliar field by occupation, study how creators organize skills, then use source context to decide what is worth opening or reusing.
01 Map a field
Browse 23 occupation groups and 867 SOC roles to learn what skills exist in adjacent domains and how they break down real work.
02 Follow creators
Use creator and repository pages to inspect maintained skill collections, recent updates, and source context before trusting a result.
03 Search with sources
Search 1.7M+ collected skills, then use occupation tags, creators, and GitHub source context to decide what is worth opening.
Start with the occupation map, then follow creators and repositories back to real code. SkillMD helps explain why a skill is worth opening, not only what it is named.
Standardizing Agent Capabilities with SKILL.md and Model Context Protocol (MCP)
In the rapidly evolving landscape of artificial intelligence, LLM agents (Large Language Model agents) have transitioned from simple text predictors to autonomous problem solvers. To orchestrate complex, multi-step agentic workflows, developers require a standardized format to specify agent capabilities, prompt instructions, system rules, and database bindings. This is where SKILL.md and the Model Context Protocol (MCP) have emerged as standard developer paradigms. SkillMD serves as the central directory for indexing, exploring, and sharing these critical agent configurations.
Our open-source registry currently tracks over 1.7 million collected SKILL.md configurations and system prompts. By compiling agent configurations from active developers on GitHub, we bridge the gap between prompt engineering research and production execution. Whether you are building agents with Anthropic's Claude Code, OpenAI's GPT-4, Google's Gemini, or local models using Ollama and LlamaIndex, standardized skill definitions ensure your agents behave predictably across different runtime environments.
What is the Model Context Protocol (MCP)?
The Model Context Protocol (MCP) is an open-source standard designed to connect LLMs to data sources, developer tools, and external environments. MCP establishes a bidirectional communication channel between client applications (like Cursor, Claude Desktop, or custom agent systems) and servers hosting data or capabilities. Standardizing instructions via SKILL.md enables LLMs to query databases, read local files, execute terminal commands, and integrate third-party APIs. SkillMD allows you to find ready-to-run MCP servers and prompt instructions for various occupations and technical tasks.
The Structure of a Professional SKILL.md File
A valid SKILL.md configuration is designed to be easily read by humans and parsed by LLMs. It contains precise system instructions, trigger conditions, required parameters, and execution examples. Below is the typical architectural blueprint of a professional agent skill:
- Metadata & Core Scope: Declares the name of the skill, author details, target models, and a description of the capability.
- Triggers & Intent Detection: Details semantic triggers that help the agent decide when to invoke this skill.
- System Prompts: Explicit system-level instructions that direct the agent's behavior, personality, safety guardrails, and formatting preferences.
- Capabilities & Tools: Lists the files, databases, or APIs the agent must access to complete the tasks.
- Few-Shot Examples: Demonstrates real inputs and outputs, helping the model generalize behavior through in-context learning.
Optimizing Agent Workflows for Modern LLMs
Writing effective agent skills requires deep knowledge of prompt engineering. With the release of advanced reasoning models like Claude 3.5 Sonnet, ChatGPT o1, and DeepSeek-V3, prompt templates must focus on structured thinking. Developers are encouraged to use XML tags (e.g., <thought>, <context>, and <rules>) to isolate execution boundaries. Standardized prompts prevent agents from suffering from context drift, ensuring that long-running tasks remain aligned with the initial system parameters.
Exploring by SOC Occupations and Creator Profiles
What makes SkillMD unique is its taxonomy. Instead of simple text search, we parse and organize files according to the Standard Occupational Classification (SOC) system. This means you can discover skills written for Computer and Mathematical roles, Business and Financial operations, Legal, Design, and and Educational Instruction fields. By tracking creator profiles, developers can study how different teams organize their custom instructions, compare version updates, and fork public configs for specialized enterprise use cases.
SkillMD operates as a high-performance index running on a fast Go backend and a highly responsive Astro SSR frontend. All search queries execute in milliseconds, featuring smart debouncing to prevent multiple API requests while keeping user data secure. Join our community of developers to standardize your AI agent instructions and optimize your LLM prompting workflows today.
Frequently Asked Questions
A practical guide to agent skills: what they are, how to inspect them, and how SkillMD helps you explore the ecosystem.