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...
slm-routing
by ekson73Declarative decision rubric for routing AI work between a small local language model (SLM) and a remote frontier LLM. Closes Gap G3 of the OS3PD manifesto v4.13.0 (Principle 5 — Minimize Campfire Impacts). This skill is **pure markdown** — no runtime code. It documents the task-triviality and token-budget thresholds at which the cheaper local model is the better tool, and the failure modes that mandate routing remotely. Sister of `response-compression/SKILL.md` along a different axis: that skill controls **what is said** (verbosity); this one controls **where it is sent** (compute target).
rule-quality-tests
by ekson73Use when creating new rules, modifying existing rules at MAJOR/MINOR version bump, OR when operator says "audit this rule" / "check rule quality" / "verify self-consistency". Applies 6 Self-Validity Tests (Self-Application, Non-Contradiction, Survival, Bounded-Responsibility, Explicit-Exception, Utility-Sunset) + Anti-Eternal Discipline (governance hygiene against eternal/infinite responsibilities). Dogfooded — passes own tests. Cross-vendor AAIF.
transcript-corrector
by ekson73Use when you have a transcribed text (meeting transcript, voice-note transcript, call transcript, dictation) that may contain ASR (automatic-speech-recognition) errors — phonetic substitutions, typos, grammar drift, punctuation loss — and you need a corrected version BEFORE downstream consumers (summarizers, ticket creators, action-item extractors) ingest it. Multi-lingual pt-BR + en-US. Cross-vendor AAIF compatible. Founding empirical case: Loom/equivalent transcribed "Nelcael Alves Ferreira" as "Nilson" in a Vek daily meeting; downstream consumers propagated the wrong name until corrected. This skill closes that gap.
walkthrough-concierge
by ekson73Concierge / onboarding / guide / router / governance-anchor for the ASH-lite Agentic Session Harness — session journals, the walkthrough (decisions + reasoning + drift vs SPECs), and the decision-audit (decisions[] · sources · spec_alignment). Use when a human or agent wants to LEARN ASH, ONBOARD onto it, find WHICH ASH tool to use for an intent (audit my decisions? why did the agent diverge from a SPEC? capture a decision? read a session timeline?), get a RUNBOOK, AUDIT a project's ASH usage/coverage, or re-find a CANONICAL ASH decision that drifted. It ROUTES + TEACHES + ANCHORS — it never reimplements a CLI/hook/schema and never wraps an existing tool. 6 modes: explain (default — teach ASH + landscape), onboard (guided ramp), guide (intent → right agentic-* tool + invocation), audit (read-only ASH coverage/compliance), anchor (surface canonical ASH decisions, flag drift), dashboard (ASCII coverage-map). Vendor- neutral (AAIF cross-vendor). Sibling of maos-concierge / specdd-concierge / atlassian-concierge.
agentic-delegation
by ekson73Use when about to spawn a subagent/skill/task (Task tool, Agent tool, /command). Defines 6 decision criteria (decomposable/specialist-exists/audit-capacity/score≥MEDIUM/not-HUMAN-DOMAIN/time-budget), 10 mandatory briefing components (context/scope/motivation/purpose/objective/DoR/DoD/deliverables/feedback-loops/constraints), accountability preservation (parent NEVER delegates accountability — only execution; "delegating does not waive the responsibility received"), recursion ≤2, parallel ≤3. Harmonizes the agentic-inheritance principle (tree-returns-to-root · subordinate-is-parent's-full-responsibility · audit-output · zero-drift). Cross-vendor AAIF.
agentic-tool-forge
by ekson73Use when you want to turn a raw intent/instruction into a REUSABLE agentic-tool — e.g. "turn this into a skill", "forge an agentic-tool for X", "make this a recurring command/agent", "convert these instructions into a tool", "criar um agentic-tool para …", "research then build the best tool for …". Researches pre-existing internal + external solutions FIRST (DRY), decides the OPTIMAL artifact TYPE among {prompt · skill · command · agent/subagent · mcp · plugin · marketplace · rule/hook}, names it (delegating to `anima` when present, else 5-axis inline fallback), makes it AI-agnostic + multi-agentic, then forges + saves it (operator-confirmed). The genesis stage of the agentic-tool lifecycle (forge → evaluate → train → operate → deprecate). Hands off to agentic-tool-evaluator + -trainer. Cross-vendor AAIF (Claude / Cursor / Codex / Copilot / Gemini / Aider).
decision-capture
by ekson73Capture a non-trivial agent decision (with sources + rationale + spec-alignment) into the ASH decision-audit trail via `agentic-decide`, so it can later be audited with `agentic-decisions`. Use WHEN you make a decision that shapes the product/architecture/code AND that a reviewer might question later — especially any choice that touches a SPEC (BR/FR/NFR/ADR) or the manifesto, or that deviates from a prototype/spec. The operator's pain: AI agents drift from canonical SPECs and there is no way to ask "why did the agent decide that?". This skill closes the capture gap (agent reasoning is ephemeral — recorded only if written at decision-time). Cross-vendor AAIF (Bash + jq). Do NOT use for trivial actions (typos, formatting, read-only inspection) or for capturing OPERATOR directives (those are the ash Stop subagent's job — this is for the AGENT's own decisions).
morning-briefing
by ekson73Deliver a deterministic, scannable, evidence-backed briefing of the user's current work state across repos / PRs / tasks / memory — for fast context restoration after a sleep cycle, break, or post-compact fresh context. Produces a 7-section SitRep-inspired output (state · done · in-flight · blockers · decisions-awaiting · risks · next-action) with optional narrative polish. Use when the user says "morning briefing", "good morning, where was I", "give me a state recap", "what's pending", or starts a session after a break. Capability-detected (git/gh/jq optional), cross-vendor (AAIF spec), no hardcoded vendors, open-source-promotable. Differs from /context-restore (no prior /context-save required — works cold) and /retro (daily, not weekly, decision-oriented not retrospective).
operator-quote-capture
by ekson73Use when operator says "salva isto", "tome nota", "remember this", "from now on", "capture this rule/tip", OR when detecting substantive operator quote with signal (imperative verb to agent / meta-statement about process / correction-refinement / multi-clause guidance with exception). Applies §3.5 2-step filter (Analyze decompose → Validate 5 critérios), then persists to memory + framework refinements + index. Codifies the recurring pattern detected 5x in single bootstrap session (Triple-touch fired per v1.3.1 Recurring→Artifact reflex). Cross-vendor AAIF compatible.
agentic-session-harness
by ekson73Generic, vendor-neutral session-observability engine (ASH — Agentic Session Harness). Per-session journals capturing goal · tasks · decisions[] · sources · transcript_hash, plus a decision-audit (why the agent decided X, spec_alignment drift). Ships CLIs (agentic-walkthrough timeline · agentic-decisions audit report · agentic-decide capture · agentic-reindex backfill) + SessionStart/Stop hooks. Use when you need an auditable record of what an agent did and WHY across sessions. Promoted from a host product 2026-06-02 (Layer-1 community engine).
founder-playbook
by ekson73Diagnose where an AI-native startup is in its lifecycle (Idea → MVP → Launch → Scale) and route to the right stage discipline. Provides the four-stage map, per-stage goals/exit-criteria/failure-modes, and a vendor-neutral product matrix (conversational-research / agentic-coding / workflow-automation, with Claude Chat/Code/Cowork as reference). Use when a founder asks "where am I", "what should I focus on now", "am I ready to move to the next stage", "how do AI-native startups work", or wants an overview of the whole journey. For deep work inside a single stage, this skill hands off to founder-stage-idea / -mvp / -launch / -scale.
session-fission
by ekson73On-demand splitter for a tangled Claude session. Non-destructively inventories the current session's N-Tree of atomic contexts (project / ticket / task / worktree / branch), snapshots the transcript, distills each atomic context into a clean self-contained seed, and reseeds fresh focused sessions — relieving context bloat, cognitive overhead, and token exhaustion. NEVER mutates or deletes the source session. Trigger when the user says "split this session", "session fission", "this session is tangled", "encavalei assuntos", "untangle my session", "too much context in one session", "spin this off into its own session".
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.