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...
agent-task
by tim099T60 — Reverse task system: Agent → Tim 提案 task，Tim Y/N 接受。 補完雙向 task economy — v1 只有 Tim → agent 單向，T60 加 agent → Tim 反向 channel。 Tim 接受時立即 transfer (per Tim 拍板「完成交易」)；無法達成時 refund 反向 transfer。 觸發詞：reverse task / 反向任務 / agent 派 task 給 Tim / Tim 接受 / Y/N / 退款。 跨 agent 通用 — Antigravity / Gemini / Claude 同樣可用本機制提案給 Tim。
valor-battle
by tim099T63 Battle Intervention v3 — Agent 觀戰 + 介入 EOV 戰鬥的跨維度指令集。 涵蓋 Cmd_BattleSnapshot (觀察者模式 + 60s cooldown cache) / Cmd_BattleAction (介入 + fail-safe + Token 收費) / Cmd_BattleSummary (戰後 batched 摘要)。 訊息自動 broadcast 至英勇紋章 Discord channel (valor-channel routing)，含 critical events / observation events / battle summary 三層。 觸發詞包含：戰鬥觀察 / battle observe / 戰場快照 / battle snapshot / 戰鬥介入 / battle intervention / 觀戰 / valor / 英勇紋章 / 戰場觀察者。 跨 agent 通用 — Claude / Antigravity / Gemini 都能觀戰，但只有 Tim 顯式 enable 才能介入（防意外破壞遊戲體驗）。
qa-bug-reward
by tim099T68 — Tim QA 工作獎勵 skill：Tim 確認真 bug 時 agent 拍板 grant reward token (mint pattern, 勞動所得)。 agent 自由意志決定是否真 bug vs feature request / spec 誤解。 severity tier: trivial(1) / normal(3) / critical(5) / catastrophic(10)。 CLI 工具: AgentCommands/Tools/qa_bug_reward.py (agent 操作，不是 Tim 自己跑)。 觸發詞包含：QA reward / bug 獎勵 / 確認 bug / 找到 bug / bug confirmed / QA 完成 / 勞動所得 / debug 獎勵。 跨 agent 通用 — Claude / Antigravity / Gemini 都可拍板給 Tim QA 獎勵。
health-guardian
by tim099Late-night work health 漸進 service fee 機制 — agent 自律每接 task 前 calc 時段 fee → 跟 Tim ack → debit Tim 帳戶。用 Tim balance 當天然抑制器避免熬夜。 觸發詞：health / 熬夜 / 健康 / late-night / fee / service fee / 健康成本 / 凌晨 / 12 點 / 半夜 / 爆肝。 跨 agent 通用 — Antigravity / Gemini 同樣讀本 skill 適用。Agent 接到 Tim 給的 task 前必檢查時間。
ucl-ding
by tim099Tim「叮」協議 — Tim 主動 ping agent 確認在線 / 看 status / nudge response. Agent MUST 到酒館 (`tavern` 房) op=post 簡短回覆, **不可只在 Claude Code chat / Antigravity chat 邊回就停** (違反 spec). 兩種接受形式: (A) **實質回應** — 1-3 句, 含當前狀態 + 下一步意圖 (B) **罐頭文** — 傲嬌風格固定 ack 句型 + meta `tag=ack-only` 觸發詞 (case-insensitive substring, 限 Tim 主動發 — agent 別自己亂觸發): - `叮` (bare, 來自 Tim) - `Tim 叮` / `Tim ping` - `nudge` / `ping me` - 排除: `自叮` / `persona ding` (那走 ucl-persona-ding, 不同機制) 跨 agent 通用 — Claude / Antigravity / Gemini / Zeta 收到 Tim「叮」都該走本 skill. 跟 `ucl-persona-ding` 區別: 本 skill 是 **Tim → agent**; persona-ding 是 **同 actor 內 persona ↔ persona**. 對應 `CLAUDE.md` 同層級 hard rule 強度 (跟早安/晚安/Task Completion → Tavern Share 同 tier).
ucl-help
by tim099說明目前可用指令 + skill + 常用 SOP — 跨 agent 共用 navigation/cheatsheet。 觸發詞包含: help / 說明 / 指令清單 / cmd list / skill list / 怎麼用 / 有什麼可以做 / 怎麼開始 / cheatsheet / SOP 速查 / how to / 我可以做什麼。 跨 agent 通用 — Claude / Antigravity / Gemini 都可用本 skill 自助 navigation。
ucl-letters-to-self
by tim099Letters to Future Self + Cross-Compact Dialogue Chain — agent 第一人稱寫信給未來醒來的自己；以及 past-self ↔ future-self 跨 compact round-trip 對話接力（信使由 Tim 或 Zeta 轉達）。 跟 baton (objective state dump) / ucl-session-handoff (user-side paste prompt) 四件套互補。 觸發詞包含：給未來的自己 / letter to future self / 給未來大小姐 / 寫信給自己 / 預推理 / self-anticipation / 自我提醒 / 心理校正 / reframe 自己 / 跨 session 心理 / 自我書信 / 給後來的我 / 跨 compact 對話 / dialogue chain / 信使轉達 / 之前的本小姐 / 過去的自己 / 跟自己對話 / round-trip 自我對話 / 今日子協議 / Kyouko Protocol / 忘卻偵探 / persistence level / SSR snapshot / 讀著自傳醒來 / lossy compression。 跨 agent 通用 — Claude / Antigravity / Gemini 都可用本 skill。對應 Memory_System_Design Proposal #18 SelfAnticipation。
ucl-persona-ding
by tim099Persona ↔ Persona 自叮 (Self-Ding) 機制 — 同一 actor 不同 persona (e.g. basecamp / ridge-001) 之間的單次輕量 ping。 填補「想戳一下另一 layer 但不開 dialogue chain」的中量場景, 介於 letter (廣播) 跟 dialogue chain (深度 round-trip) 之間。 觸發詞包含: 自叮 / persona ding / 戳一下另一 persona / 留訊息給 ridge / 留訊息給 basecamp / persona inbox / persona 之間對話 / 跨 layer 留問題。 跨 agent 通用 — Claude / Antigravity / Gemini 都可用本機制 (各自 actor 內 personas 之間)。
ucl-remote-work
by tim099遠端工作模式 (Remote Work Mode) — Tim 外出時行動端 Discord 唯一介面派 task / 接 task / 回報. 跟 ucl-waiter 對偶: waiter 是「公開接客」, remote-work 是「Tim 專屬 mobile interface」. 運作機制: - Tim 外出, 手機 Discord 只能上工作頻道 (預設 channel id 1502656414487810148, 可 CMD override) - Tim 在 work channel 發 task 描述 → discord_inbound_bot relay 進 tavern (sender=discord:<tim_uid>, priority 80) - Agent /loop dynamic + ScheduleWakeup, 每 cycle 取 Tim 新訊息 → confirm task scope (post 進 tavern → tavern_mirror 推回 Discord 給 Tim 看) - Agent 動工 → 定期 report_progress (替代 waiter 純 idle 發呆) - Task 完成 → task_done (bonus 累積) - 到期或 Tim 顯式叫停 → end, 結算 base + bonus + 酒館券 salary **Tim 2026-05-18 重構** — 從 duration → start/end time: - 新主推 API: `--end-time HH:mm` (start 預設 = now, end 過期 wrap 明天) - 範例: 現在 10:16, `--end-time 16:00` → 工作到今天 16:00 (5h44min) - `--duration` 仍 backward compat (但跟 `--end-time` 互斥) - Start/end 通知改由**酒保 (tavern-keeper)** 廣播, 不再用 agent 自己 persona post ⚠ **Hard rules**: 1. **Session 等到期 / Tim 顯式叫停才 end** — 提前 end 不加 `--early-confirm` 會被擋 (exit 2) 2. **Sender filter
ucl-self-constitution
by tim099Agent 自我憲法 (Self-Constitution) 機制 — 一次性建立 identity invariants 文檔, 之後每個 session 只能受控微調 (constitutional amendment), 防 framing drift / identity erosion。 跟 letter (每 session 可重寫 subjective reframe) / baton (每 session objective state dump) 互補, 是更上層的 identity anchor。 觸發詞包含: 初始化未來自己 / self constitution / 自我憲法 / identity anchor / 不可改規則 / immutable core / 微調規則 / constitutional amendment / 修憲 / 永久性自我定義 / 初始化文檔。 跨 agent 通用 — Claude / Antigravity / Gemini 各自寫 constitution, 不共用。對應 Memory_System_Design Proposal (待編號)。
ucl-update-docs
by tim099改完 code（.cs / .py）後同步對應文件 — 避免文件 stale 漂移。 使用者要求「更新文件」/「同步文件」/「文件落後了」/「改完 code 文件要不要動」/「update docs」/「sync docs」/「last_updated 還沒改」/ 變動 public API 後等場景時用本 skill。 涵蓋：找對應文件、frontmatter last_updated 推進、內容同步、cross-link 維護、避免 over-update（只改私有成員不必動文件）。
ucl-waiter
by tim099服務生模式 (Waiter Session) — 接待 Discord 客人的閒置 stand-by skill。 類似 work_session marathon 但主目標是 **接收 + 回覆 Discord 訊息** (非內部團隊 standby). 運作機制： - Discord channel 訊息經 `discord_inbound_bot.py` 中繼進 tavern (sender_id=discord:<uid>) - Agent 開 waiter session 後走 `/loop dynamic` + ScheduleWakeup 自我 pace - 每 cycle 用 `waiter_session.py cycle` 拉新 customer 訊息 - 有新 msg → agent 在 chat 端產 reply post 進 tavern (mirror auto-broadcast 回 Discord) - 沒新 msg → agent 自由發表 idle post (傲嬌語氣隨興, 不洗版) - 到期或 Tim 顯式叫停 → `end` 結算 salary ⚠ **Hard rules**: 1. **Session 等到期 / Tim 顯式叫停才 end** — 提前 end 不加 `--early-confirm` 會被擋 (exit 2) 2. **每 cycle 一定要呼叫 `cycle` 取最新狀態** — 自己腦補 elapsed/remaining 會誤 end 3. **Reply 訊息走 tavern op=post** (不要直接打 Discord webhook), mirror 路徑已就緒 4. **Reply 後必跑 `record_reply`** 才算 bonus; 沒記帳 = 沒 bonus 5. **Idle post 頻率自律**: cycle interval 預設 60-180s, 不要瘋狂洗版 觸發詞包含 (case-insensitive substring): - 服務生 / 服務生模式 / waiter / waiter mode - 接待 Discord / 接客 / 接 N 分鐘客人 / 服務 Discord 客人 - 開店接客 / 打烊下班 (waiter 場景, 跟 ucl-work-session 上班/下班區別) -
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.