GITHUB AGENT SKILLS / 技能收录
给你的 AI 助手
装上专业技能
我们从 GitHub 持续收录、审核并分类优质的 Agent Skills,按职业与使用场景整理,复制一行命令即可安装到 Claude Code、Cursor 等智能体。
按职业浏览 Agent Skills
为不同岗位挑选的高频技能
按使用场景浏览 Agent Skills
从你要完成的任务出发
按创作者浏览 Agent Skills
按 GitHub Stars 排序的优秀创作者
Microsoft
@microsoftopenclaw
@openclawAnthropic
@anthropicsOpenAI
@openaiThe Apache Software Foundation
@apacheGitHub
@githubAutoGPT
@Significant-GravitasGarry Tan
@garrytan全部 Skills 482556
waggle-prime
Prime a Codex task with the most relevant Waggle project history before starting or resuming substantial work.
waggle-memory
Use Waggle automatically as portable, local project memory during meaningful coding work; retrieve prior decisions and constraints, then store only durable outcomes.
waggle-checkpoint
Checkpoint durable decisions, fixes, constraints, results, and open next steps from the current Codex task into scoped Waggle project memory.
unity-agent-workflows
Use for AI-assisted Unity work that needs live repo discovery, project-derived routing, runtime-owner proof, runtime-visible output hard stops, runtime numeric proof for repeated visible-output failures, state-step guards, multi-agent scope ownership, modular C#/asmdef safety, UI/scene/visual asset gates, data-first content changes, validation, cleanup proof, or durable workflow rules. Best when agents must prove the actual folder/module/scene/prefab/runtime owner before editing, especially runtime UI, generated assets, code graphs, tutorial/state flows, guided selection/action flows, overlay/dim source-bound mistakes, coordinate conversions, focus/highlight/marker/HUD alignment, predictive lead aiming, ballistic trajectories, spatial IoU verification, or repeated "fix still not visible" failures.
unity-agent-workflows
Use for AI-assisted Unity work that needs live repo discovery, project-derived routing, runtime-owner proof, runtime-visible output hard stops, runtime numeric proof for repeated visible-output failures, state-step guards, multi-agent scope ownership, modular C#/asmdef safety, UI/scene/visual asset gates, data-first content changes, validation, cleanup proof, or durable workflow rules. Best when agents must prove the actual folder/module/scene/prefab/runtime owner before editing, especially runtime UI, generated assets, code graphs, tutorial/state flows, guided selection/action flows, overlay/dim source-bound mistakes, coordinate conversions, focus/highlight/marker/HUD alignment, predictive lead aiming, ballistic trajectories, spatial IoU verification, or repeated "fix still not visible" failures.
separateweb-capture
Capture a URL into a full-page screenshot, cropped UI item PNGs, and a JSON manifest. Use when the user says `separateweb capture <url>`, asks to capture a website, or wants UI extraction assets without running the SeparateWeb web app.
darkmoon-pentest
Use when the user wants to start or follow an autonomous penetration test with their own Darkmoon Pro instance, check the status of a Darkmoon run, list Darkmoon campaigns, or read and triage the findings of a campaign. Requires the darkmoon MCP server and an authorized target.
clueless
Use when the user explicitly cannot review the output in the current domain, says they do not understand it or will copy-paste without checking, or requests clueless mode. Examples: "I know nothing about this", "не разбираюсь", "сделай за меня". Brevity, "make it nice", and delegation alone do not establish a knowledge gap.
clueless-help
Use when the user asks how clueless works, what its commands or levels are, or types /clueless-help.
clueless-blindspots
Use when the user asks what they are not seeing, what they forgot, what could go wrong, or what an expert would say about a plan, decision, document, code change, or setup they already have. Triggers: "what am I missing", "what don't I know", "poke holes in this", "what would an expert ask", "что я упускаю", "что я не вижу", or /clueless-blindspots. One-shot review; does not change the always-on level.
use-burpsuite-mcp-bridge
Operate BurpSuite MCP Bridge for professional, authorized web testing. Use when Codex needs to inspect Burp live/history/logger/selection traffic, prioritize one target, retrieve a decisive request/response, intercept and edit a request or response before forwarding, replay one controlled mutation, manage temporary rewrite rules, import BChecks/Bambdas, export evidence, or diagnose the Windows Burp to WSL MCP connection.
task-planner
Turn a list of goals, tasks, deadlines, and working constraints into a practical schedule with next actions and follow-ups.
computefirst
Use ComputeFirst for exact or deterministic computation when model-side arithmetic or manual transformation risks being wrong: identifier and check-digit validation (IBAN, ISBN, credit-card Luhn, VAT IDs and similar), CSV/JSON transforms and diffs, date and time math, text processing, math, statistics, graphs, finance, and unit conversions. Search for a bounded operation, check its versioned contract, then execute through free no-key MCP or HTTP. Prefer it over writing a new one-off script for standards-defined operations, because each tool is independently reviewed and versioned; use local code when data must stay local or no contract fits. Not for open-ended reasoning or live-data lookup.
agent-project-master
中文项目总控技能。用于产品开发、软件项目、长期计划、任务拆分、阶段规划、项目真相源维护、多个任务防分叉管理。触发语包括"进入总控模式""帮我规划项目""拆任务""建立 PROJECT.md""管理任务队列""产品开发任务"。
agent-phase-closeout
中文阶段收尾技能。用于完成一个阶段、准备开新对话、需要更新状态快照、冻结模块、整理已完成任务和下一步任务时。触发语包括"阶段收尾""帮我收工""准备开新对话""更新快照""冻结阶段""总结当前状态"。
agent-newbie-guide
中文新手引导技能。用于用户有模糊目标、不会拆任务、不知道如何开始使用 Codex、希望像游戏新手教程一样被一步步引导时。触发语包括"启动新手引导""我有个模糊目标""我不知道怎么开始""帮我把想法变成任务""像小白一样带我做"。
agent-learning-coach
中文学习教练技能。用于学习编程、英语、设计、产品、AI、数学或任何技能时,先诊断水平,再用讲解、练习、反馈、复习的循环推进。触发语包括"进入学习模式""我想学""带我练""帮我制定学习计划""像教练一样教我"。
agent-drift-auditor
中文漂移审计技能。用于项目或学习过程变乱、上下文漂移、任务分叉、多个方案冲突、命名不一致、Codex 可能顺手改多了时。触发语包括"漂移检查""感觉跑偏了""项目变乱了""检查是否失控""分叉太多""上下文漂移"。