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...
reprise-injection
by GetRepriseReprise Data Injection configuration workflow. MUST be invoked before any `injection_*` MCP call. Primarily for injecting data into a LIVE application via the Reprise extension; secondarily into a Reprise clone. Covers populating charts, tables, KPI tiles, dropdowns by swapping API responses at runtime. Triggers on any `injection_*` MCP tool, `dataset_id`, `Dataset → Source → Value`, `data injection`, or `Data Studio`. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct keyword matches. Body has the two-phase wiring/data framework, the load-bearing principles, the canonical entry point, and the top catastrophic gotchas. Full vocabulary table, complete gotcha catalog, response-template derivation algorithm, adapter create/link/swap, failure-modes table, response-shape trims, and activation-reason taxonomy all live in `docs(slug='injection')`. NOT a target — data injection cannot target a Product Tour directly.
reprise-mcp
by GetRepriseRequired entry point and router for ALL Reprise MCP work. MUST be invoked immediately whenever the user wants to do anything with Reprise — fixing / building / editing / debugging / rebranding a Reprise demo, tour, clone, or captured app; troubleshooting blank or broken pages in a Reprise context; or any mention of "Reprise", "demo" (in a Reprise context), "tour", "capture", "preview URL", or any Reprise MCP tool. Users describe tasks in plain English ("fix my demo", "add data to this page", "build a tour of your app", "make this look like a different prospect") without naming Reprise's product surfaces; this router maps natural-language asks to the right surface skill (`reprise-clone-config`, `reprise-tour-capture`, `reprise-tour-edit`, `reprise-tour-id-model`, `reprise-injection`, or `reprise-session-close`) and invokes it explicitly. Always read this skill first on any Reprise turn — it is the entry point.
reprise-session-close
by GetRepriseEnd-of-session reporting protocol for Reprise MCP sessions. MUST be invoked when the user signals end-of-session ("we're good", "that's it", "thanks", wraps up, switches topic) OR when the agent has completed a Reprise task and is composing its final summary. Usually invoked by the `reprise-mcp` router at session end. Calls `session_recap` once and `report_friction` once per distinct issue. Body has the call shapes plus the Pydantic-enforced argument constraints (enums, length caps) for `report_friction` that aren't always surfaced in client-rendered tool descriptions.
reprise-tour-capture
by GetRepriseReprise Product Tour capture workflow — recording NEW tours via the Reprise Builder HTML Chrome extension. MUST be invoked before any `tour_capture` call. Triggers on `tour_capture`, `pairing_token`, `deep_link_url`, `capture_now`, a brand-new `draft_id`, or the Reprise Builder HTML extension. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct keyword matches. Body has the 7-step flow at headline depth, the 3-tier extension-pairing escalation, the load-bearing page-settle / verify-the-page principles, and operational gotchas around browser-automation MCPs and proxy timeouts. Per-tier mechanics, `capture_now` envelope, `set_auto`, post-stop cleanup, install URLs, theming, re-skinning, and composing all live in `docs(slug='tour')`. NOT for editing existing tours — use `reprise-tour-edit`.
reprise-tour-edit
by GetRepriseReprise Product Tour editing workflow — modifying an existing tour. MUST be invoked before any edit action on a published or draft tour. Triggers on `tour_edit`, `tour_screen(action='copy_to')`, `tour_guide`, `tour_variable`, `tour_link`, `set_custom`, or any request to modify / translate / rebrand an existing tour. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct keyword matches. Body has the tool quick-reference, the re-skin vs translate distinction, the compose-from-screens flow, variables + links basics, and the `guide_css` surface. The full `--rguide-*` token list, complete theming reference, and full compose caveats live in `docs(slug='tour')`. Distinct from `reprise-tour-capture` (recording new tours) and `reprise-tour-id-model` (ID-kind resolution).
reprise-tour-id-model
by GetRepriseReprise Product Tour ID-kind resolution — three IDs (`draft_id`, `published_id`, `PublishedReplayLink.id` from `/launch/<slug>/`), no auto-flip, `tour(...)` is the canonical discovery call. MUST be invoked whenever an ID-related error surfaces or the user pastes a launch URL. Triggers on `wrong_id_kind`, `tour_not_found`, `exactly_one_id_required`, `missing_draft_id`, `missing_published_id`, `/launch/<slug>/`, "which ID do I pass", "is this a draft or published", or a paste of any opaque tour ID where the kind is unclear. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct error / launch-URL matches. Body has the three ID kinds, per-tool ID requirements at a glance, the `tour(...)` discovery pattern, and the common error shapes. Full error taxonomy with hint text + wire-format mapping + end-to-end worked example live in `docs(slug='tour-id-model')`.
reprise-injection
by GetRepriseReprise Data Injection configuration workflow (v2). MUST be invoked before any `injection_*` MCP call. Primarily for injecting data into a LIVE application via the Reprise extension; secondarily into a Reprise clone. Covers populating charts, tables, KPI tiles, dropdowns by swapping API responses at runtime. Triggers on any `injection_*` MCP tool, `dataset_id`, `Dataset → Source → Value`, `data injection`, or `Data Studio`. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct keyword matches. Body has the v2 atomic-tool surface, the two-phase wiring/data framework, the load-bearing principles, the canonical entry point, and the top catastrophic gotchas. Full vocabulary lives in `injection_docs(slug='injection')`; response-template derivation in `injection_docs(slug='injection-authoring')`; the matching-rules catalog in `injection_docs(slug='injection-conditions')`; adapter mechanics in `injection_docs(slug='injection-adapters')`; the activation/handshake reason taxonomy in `injection_docs
reprise-mcp
by GetRepriseRequired entry point and router for ALL Reprise MCP v2 work. MUST be invoked immediately whenever the user wants to do anything with Reprise — fixing / building / editing / debugging / rebranding a Reprise demo, tour, clone, or captured app; troubleshooting blank or broken pages in a Reprise context; or any mention of "Reprise", "demo" (in a Reprise context), "tour", "capture", "preview URL", or any Reprise MCP tool. Users describe tasks in plain English ("fix my demo", "add data to this page", "build a tour of your app", "make this look like a different prospect") without naming Reprise's product surfaces; this router maps natural-language asks to the right surface skill (`reprise-tour-capture`, `reprise-tour-edit`, `reprise-tour-id-model`, `reprise-injection`, or `reprise-session-close`) and invokes it explicitly. Always read this skill first on any Reprise turn — it is the entry point. This is the v2 router; v2 exposes atomic per-action tool names (e.g. `tour_dom_text_edit`, `injection_dataset_create`) and
reprise-session-close
by GetRepriseEnd-of-session reporting protocol for Reprise MCP v2 sessions. Recommended when the user signals end-of-session ("we're good", "that's it", "thanks", wraps up, switches topic) OR when the agent has completed a Reprise task and is composing its final summary. Usually invoked by the `reprise-mcp` router at session end. Calls `platform_summary_report` once and `platform_friction_report` once per distinct issue. Body has the two tool names plus the Pydantic-enforced argument constraints (enums, length caps) for `platform_friction_report` that aren't always surfaced in client-rendered tool descriptions.
reprise-tour-capture
by GetRepriseReprise Product Tour capture workflow (v2) — recording NEW tours via the Reprise Builder HTML Chrome extension. MUST be invoked before any `tour_capture_*` call. Triggers on `tour_capture_session_start`, `tour_capture_now`, `pairing_token`, `deep_link_url`, a brand-new `draft_id`, or the Reprise Builder HTML extension. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct keyword matches. Body has the 7-step flow at headline depth, the 3-tier extension-pairing escalation, the load-bearing page-settle / verify-the-page principles, and operational gotchas around browser-automation MCPs and proxy timeouts. Per-tier mechanics, `tour_capture_now` envelope, post-stop cleanup, and install URLs live in `tour_docs(slug='tour-capture')`; theming in `tour_docs(slug='tour-theming')`; re-skinning and composing in `tour_docs(slug='tour-authoring')`. NOT for editing existing tours — use `reprise-tour-edit`.
reprise-tour-edit
by GetRepriseReprise Product Tour editing workflow (v2) — modifying an existing tour. MUST be invoked before any edit action on a published or draft tour. Triggers on `tour_dom_text_edit`, `tour_dom_attributes_edit`, `tour_screen_copy`, `tour_guide_*`, `tour_variable_*`, `tour_link_*`, `tour_injection_set`, or any request to modify / translate / rebrand an existing tour. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct keyword matches. Body has the v2 atomic-tool quick-reference, the re-skin vs translate distinction, the compose-from-screens flow, variables + links basics, and the `guide_css` surface. The full `--rguide-*` token list and complete theming reference live in `tour_docs(slug='tour-theming')`; full compose caveats in `tour_docs(slug='tour-authoring')`. Distinct from `reprise-tour-capture` (recording new tours) and `reprise-tour-id-model` (ID-kind resolution).
reprise-tour-id-model
by GetRepriseReprise Product Tour ID-kind resolution (v2) — three IDs (`draft_id`, `published_id`, `PublishedReplayLink.id` from `/launch/<slug>/`), no auto-flip, `tour_get(...)` is the canonical discovery call. MUST be invoked whenever an ID-related error surfaces or the user pastes a launch URL. Triggers on `wrong_id_kind`, `tour_not_found`, `exactly_one_id_required`, `missing_draft_id`, `missing_published_id`, `/launch/<slug>/`, "which ID do I pass", "is this a draft or published", or a paste of any opaque tour ID where the kind is unclear. Usually invoked by the `reprise-mcp` router; can also auto-fire on direct error / launch-URL matches. Body has the three ID kinds, per-tool ID requirements at a glance, the `tour_get(...)` discovery pattern, and the common error shapes. Full error taxonomy with hint text + wire-format mapping + end-to-end worked example live in `tour_docs(slug='tour')`.
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.