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...
js-regex-cjk-word-boundary
by divinevideoFix JavaScript regex failures when matching CJK (Chinese/Japanese/Korean) text using \b word boundaries. Use when: (1) Regex pattern with \b silently fails to match Japanese, Chinese, or Korean text, (2) Pattern works for Latin/ASCII text but not CJK, (3) hasDriverLicenceCue or similar text-detection function returns false for CJK input despite correct characters, (4) inferIssuerFromTitleText or pattern-matching functions fail on non-Latin scripts. Root cause: JavaScript \b only recognizes [a-zA-Z0-9_] as "word characters" — CJK characters are classified as \W (non-word), so \b before/after CJK always sees a non-word/non-word boundary and fails to match.
clickhouse-rust-type-mismatches
by divinevideoFix ClickHouse query errors in Rust when using clickhouse-rs crate. Use when: (1) "string is not valid utf8" errors - typically FixedString columns need CAST to String, (2) "tag for enum is not valid" errors - typically Option<T> fields receiving non-NULL values, (3) Sum/count aggregations returning Float64 but Rust expects u64, (4) LEFT JOIN results with empty strings where Rust expects Option::None, (5) "not enough data, probably a row type mismatches a database schema" - query SELECT returns fewer columns than the Rust Row struct expects (common when multiple query functions share the same struct but one function is missing columns added later), (6) Garbled/corrupted data when INSERTing to FixedString(N) columns using String Rust type - silently corrupts data (no error!) because String adds a length prefix but FixedString expects exactly N raw bytes. Fix: use [u8; N] with #[serde(with = "BigArray")] from serde-big-array, (7) "the trait bound `[u8; 64]: serde::Serialize` is not satisfied" when using [u8; N]
nostr-addressable-event-dual-tag-query
by divinevideoFix "count shows X but list shows empty" bugs when querying related events (comments, reactions, zaps, reposts) for Nostr addressable events (Kind 30000-39999). Use when: (1) REST API shows engagement count > 0 but WebSocket query returns empty, (2) Comments/reactions exist but app shows "No comments yet" or similar, (3) Working with NIP-22 comments, NIP-25 reactions, or NIP-18 reposts on Kind 34235/34236 video events or other addressable events. Root cause: clients may reference addressable events using either E tag (event ID) or A tag (kind:pubkey:d-tag format), and querying only one tag misses events tagged with the other.
nostr-addressable-event-d-tag-requirement
by divinevideoFix duplicate Nostr events (Kind 30000+ parameterized replaceable events) caused by missing d-tag. Use when: (1) Same content appears multiple times on relay for same pubkey, (2) Events aren't being replaced/updated as expected, (3) Publishing NIP-71 video events (Kind 34236) or other addressable events and seeing duplicates. The d-tag is REQUIRED for addressability - without it, each publish creates a new non-replaceable event.
fastly-compute-publish-ignores-draft-versions
by divinevideoFix "backend exists on a draft version but isn't reachable from the new Compute deploy" bugs after `fastly compute publish`. Use when: (1) You ran `fastly service backend create --autoclone` (or any dashboard/CLI change that creates a draft version) but did not activate that draft, (2) You then ran `fastly compute publish` which appeared successful but your backend / ACL / header / dictionary change is missing from the active version, (3) Compute code returns backend-not-found errors or misroutes requests, (4) You're surprised to see your draft version number "skipped over" in the version chain. Root cause: `fastly compute publish` clones from the currently-active version, not from the latest draft, so dashboard changes made in an un-activated draft get stranded.
art-direct
by divinevideoArt direction for any content — reads text, PDF, Word, HTML, PPT, then proposes 2-3 creative directions with photography style, mood, and visual language. After selection, generates AI image prompts and visual briefs section-by-section. Use when the user shares content and needs visual direction, image sourcing, or creative direction for any material.
mockito-stale-mock-silent-trycatch-failure
by divinevideoFix misleading Flutter/Dart test failures where expected data is empty ([]) or default values instead of the mocked response, caused by stale Mockito generated mocks or missing stubs being silently swallowed by try/catch in provider/repository code. Use when: (1) Test expects non-empty data but gets [], (2) Mock stubs for a method are set up but the code under test returns fallback/empty results, (3) A PR added new methods to a service class and tests that previously passed now fail with misleading "expected X, actual: empty" assertions, (4) "Generated files are out of date" CI error alongside test failures. The root cause is MissingStubError thrown by unstubbed mock methods, caught by production try/catch blocks, causing silent fallback to empty state.
nostr-rest-api-field-mapping-gap
by divinevideoFix features broken by REST API responses that flatten Nostr events, losing tag data. Use when: (1) A feature works for WebSocket-loaded events but not REST API-loaded ones, (2) A model field (like sha256, textTrackRef) is null for REST API data but set for WebSocket data, (3) REST API returns denormalized fields (d_tag, video_url) but omits raw tags array, (4) hasSubtitles/hasFeature returns false for REST-loaded videos. Common in Nostr clients that use both REST APIs (for analytics/bulk queries) and WebSocket (for real-time events).
nostr-replaceable-event-mutation-overwrite
by divinevideoFix silent data loss when mutating Nostr replaceable events (Kind 0 profile, Kind 3 contact/follow list, Kind 10002 relay list, etc.) in client apps. Use when: (1) Following someone wipes the user's entire follow list, (2) Updating profile metadata loses existing fields, (3) Fresh browser session or mobile login causes data loss on first action, (4) Replaceable event mutation uses stale or null cached state. Root cause: Nostr replaceable events are full-replace (no partial update), so publishing based on stale/unloaded cache overwrites the canonical version. Applies to any Nostr client using React, Flutter, or similar reactive frameworks where query state may not be loaded when a mutation fires.
e2e-test
by divinevideoRun and debug Flutter E2E integration tests that exercise the real app against a local Docker backend (no mocks). Use when running E2E tests, debugging failures, or working on the local harness.
figma-designer
by divinevideoAnalyze Figma designs and translate them into Flutter code using existing divine_ui components and VineTheme. Use when implementing UI from Figma mockups or when checking if a component already exists before building new. Invoke with /figma-designer.
gcloud-builds-tag-deploy-false-success
by divinevideoFix false-positive Cloud Run deploys where `gcloud builds submit --tag` silently fails but the follow-up `gcloud run services update --image :latest` succeeds against a stale image, producing a confident "Deploying... Done" message with pre-fix code still live. Use when: (1) gcloud builds submit prints "ERROR: (gcloud.builds.submit) Invalid value for [source]: Dockerfile required when specifying --tag" but the shell scrolls past it, (2) code changes verified locally aren't visible in production after a deploy that printed "revision ... has been deployed and is serving 100 percent of traffic", (3) Chaining `gcloud builds submit` and `gcloud run services update` in a single copy-paste block, (4) Using the mutable `:latest` tag for Cloud Run deploys.
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.