# Skills Radar

> Discover, compare, and install Agent Skills from across the ecosystem (GitHub, skills.sh, skillregistry, npm, Hacker News, Dev.to, awesome lists). Use at the START of a task to pull the relevant existing capabilities and a same-domain comparison instead of researching the web — e.g. "what skills exist for AI video / PDF / testing / deployment", "compare similar skills", or "find and install a skill for X".

- Skill: `tomtdhzz/skills-radar` (Agent Skill, multi-file: 11 files)
- Install (CLI): `npx skillmds@latest add tomtdhzz/skills-radar`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomtdhzz/skills-radar/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: tomtdhzz (https://skillmd.com/u/tomtdhzz)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomtdhzz/skills-radar

---


# skills-radar

A multi-source knowledge base of Agent Skills. Use it to answer "what capabilities
already exist for this goal, which is best, and how do I install it" — without manual
web research.

## When to use
- The user starts a new task/business ("I want to build an AI video product") and you
  want the relevant existing skills up front.
- The user asks to find, compare, or install a skill for some domain.
- Before decomposing/implementing, to prefer reuse over building from scratch.

## How to use
Run from the skills-radar project directory (clone: `github.com/<you>/skills-radar`):

```bash
# Capability lookup for a goal (Chinese or English). Prints matched domains
# + a comparison table: sources / installs / stars / score / install command.
python3 kb.py query "<the user's goal or domain keywords>"

# Keyword search across the catalog.
python3 crawl.py search "<keyword>"

# Refresh the catalog + vault (also runs daily via launchd/CI).
python3 crawl.py crawl
```

Present the returned candidates to the user, let them choose, then install:

```bash
npx skills add <owner/repo/skill> -g -y     # skills.sh / github / awesome
npm i <package>                              # npm-distributed
sr install <id>                              # skillregistry
# or:
python3 crawl.py pull <id>
```

## Notes
- Ranking favors skills confirmed by multiple sources (cross-source score), then popularity.
- Output is a recommendation to verify, not an automatic install; confirm before depending on it.
- Auto-install is gated behind a whitelist (`crawl.py pull --auto`, off by default).
- Full docs: README.md. Obsidian vault export + workflow integration: `docs/`.