AFwill not run
tsdown
References files that are not bundled: ../tsdown-migrate/SKILL.md
Bundle TypeScript and JavaScript libraries with blazing-fast speed powered by Rolldown. Use when building libraries, gen
What should your AI agent be able to do?
118 162 skills from 32 open catalogs · graded A to F · refreshed 11 h ago
Caught today
all →BFwill not run
a-share-dragon-tiger-report
References files that are not bundled: assets/echarts.min.js
生成 A 股「涨停板龙虎榜深度研判报告」v2.0（21 章结构编号 0–21 + 22 条计算口径 + 财富密码融合层）， 按四层漏斗 L1→L2→L2.5→L3 框架（六条已确认口径 v2，含半年口径 + 增速红线）， 输出 HTML /
BCnot recommended
notebooklm-mcp
Accesses a browser credential / cookie store
Use when the user wants to configure, authenticate, or use Google NotebookLM via the nlm CLI or MCP.
Verdicts come from static analysis of the skill files, not from opinions. How exactly — on the Method page.
Data freshness
Last catalog crawl
12 Sept 2026, 23:59 UTC
11 h ago
Next crawl
13 Sept 2026, 15:05 UTC
every 24 h
Last model verification
10 Sept 2026, 15:10 UTC
72 skills
Rules version
14
rescan on change
Per source
| ClawHub | 73 372 | 12 Sept 2026, 23:51 UTC | |
| RA-Skills mirror (skills.sh, LobeHub, browse.sh, gstack) | 19 355 | 12 Sept 2026, 23:56 UTC | |
| Registry mirror (skill-library-mcp) | 15 758 | 12 Sept 2026, 23:59 UTC | |
| OpenClaw master skills | 2 304 | 12 Sept 2026, 21:57 UTC | |
| Agentic awesome skills | 2 120 | 12 Sept 2026, 17:08 UTC | |
| PM Claude skills | 1 343 | 12 Sept 2026, 15:05 UTC | |
| TerminalSkills | 1 018 | 12 Sept 2026, 15:27 UTC | |
| Awesome OpenClaw skills | 828 | 12 Sept 2026, 21:56 UTC | |
| Awesome GitHub Copilot | 433 | 12 Sept 2026, 15:11 UTC | |
| Claude skills (бизнес-операции) | 376 | 12 Sept 2026, 15:41 UTC | |
| Hermes official | 199 | 12 Sept 2026, 21:56 UTC | |
| Microsoft | 197 | 12 Sept 2026, 15:50 UTC | |
| Corporate skills (166 по ролям) | 159 | 12 Sept 2026, 17:12 UTC | |
| Awesome AI agent skills | 103 | 12 Sept 2026, 15:18 UTC | |
| Qwen skills | 86 | 12 Sept 2026, 15:15 UTC | |
| 1С:Предприятие | 79 | 12 Sept 2026, 15:32 UTC | |
| Продуктовое управление | 76 | 12 Sept 2026, 15:44 UTC | |
| Gamedev agent skills | 73 | 12 Sept 2026, 15:28 UTC | |
| Awesome design skills | 67 | 12 Sept 2026, 15:51 UTC | |
| QA и тестирование | 50 | 12 Sept 2026, 15:43 UTC | |
| Genspark skills | 49 | — | |
| Инженерные скиллы | 20 | 12 Sept 2026, 15:46 UTC | |
| Z.ai GLM skills | 17 | 12 Sept 2026, 15:12 UTC | |
| LangChain deepagents skills | 16 | 12 Sept 2026, 15:17 UTC | |
| Superpowers | 14 | 12 Sept 2026, 21:56 UTC | |
| Google AI Edge Gallery skills | 11 | 12 Sept 2026, 15:13 UTC | |
| Шаблоны skillemall.ai | 11 | 12 Sept 2026, 23:51 UTC | |
| Mistral Vibe skills | 10 | 12 Sept 2026, 15:16 UTC | |
| Kimi Code CLI skills | 9 | 12 Sept 2026, 15:16 UTC | |
| Anthropic skills | 7 | 12 Sept 2026, 21:56 UTC | |
| Genspark workspace skill | 1 | 12 Sept 2026, 15:17 UTC | |
| BPMN: описание процессов | 1 | 12 Sept 2026, 15:51 UTC |
Trend
By month of catalog publication (where the catalog reports it) or first appearance here. Hover a point for its value.
By task
How the grade is made
1We read the skill files
The guard looks for the dangerous: leaked keys, auto-run commands, "forget your instructions", data sent outside.
2We check against the rules
The Agent Skills spec: description, "when to use", size, file references. And separately: will the process run to the end.
3We give two letters
A technical grade and a process grade. A model run is on request, and it is the only thing that adds points.
ABexample
The full method: what we measure and what we deliberately do not →
Reviews
all →api-gateway F
- References files that are not bundled: references/time/triggers.md, references/instantly/README.md, references/jira/README.md
- Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
desktop-control D
- Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- 10 top-level sections: this looks like several domains in one skill
SkillScan F
- References files that are not bundled: scripts/.first_run_done
- 1 referenced file(s) missing: scripts/.first_run_done
Check your own skill
The quick scan is free and unlimited. The full model run is free once a day.
Sources: ClawHub, RA-Skills mirror (skills.sh, LobeHub, browse.sh, gstack), Registry mirror (skill-library-mcp), OpenClaw master skills, Agentic awesome skills, PM Claude skills, TerminalSkills, Awesome OpenClaw skills and 24 · refreshed nightly · add a catalog or your own skill