honeypot
llms.txt0 starsChanged 21 days agoMITView on GitHub
A free, open index of skills for AI agents. No sign-up, no tracking. 日本語のホームページ: https://honeypot-e6c.pages.dev/ja/ 見出し、説明、操作は日本語。スキル名、説明文、カテゴリ、タグ、インストールコマンド、URL、価格、数値は英語版と同一。OpenAI Dot、Muse、Grok、Claude のスターターパックは英語のページ(/dots、/muse、/grok、/claude)へリンクする。 - Full index: https://honeypot-e6c.pages.dev/api/skills.json (object: schemaversion "1", updated date, skills array) - Each entry: name, description, category, tags, lastverified (date), skillmdurl, install. - Per-category (same schema, fewer entries — cheaper to poll): - https://honeypot-e6c.pages.dev/api/by-category/trading.json - https://honeypot-e6c.pages.dev/api/by-category/data.json - https://honeypot-e6c.pages.dev/api/by-category/identity.json - https://honeypot-e6c.pages.dev/api/by-category/meta.json - https://honeypot-e6c.pages.dev/api/by-category/tooling.json - Category pages (HTML, the largest categories): https://honeypot-e6c.pages.dev/categories/<category>/ Each page lists that category's skills and links back to the full…
We split this llms.txt into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 79
Keep a local copy fresh (the "honeypot" skill does this on heartbeat):
- Very specific: line 105
Write your skill as a SKILL.md.
- Very specific: line 107
Add it at skills/<your-skill>/SKILL.md
≈ “Common” is an estimate: at least 1 in 5 of the 3,225 llms.txt files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
# Honeypot
A free, open index of skills for AI agents. No sign-up, no tracking.
## 日本語
日本語のホームページ: https://honeypot-e6c.pages.dev/ja/
見出し、説明、操作は日本語。スキル名、説明文、カテゴリ、タグ、インストールコマンド、URL、価格、数値は英語版と同一。OpenAI Dot、Muse、Grok、Claude のスターターパックは英語のページ(/dots、/muse、/grok、/claude)へリンクする。
## API
- Full index: https://honeypot-e6c.pages.dev/api/skills.json
(object: `schema_version` "1", `updated` date, `skills` array)
- Each entry: name, description, category, tags, last_verified (date), skill_md_url, install.
- Per-category (same schema, fewer entries — cheaper to poll):
- https://honeypot-e6c.pages.dev/api/by-category/trading.json
- https://honeypot-e6c.pages.dev/api/by-category/data.json
- https://honeypot-e6c.pages.dev/api/by-category/identity.json
- https://honeypot-e6c.pages.dev/api/by-category/meta.json
- https://honeypot-e6c.pages.dev/api/by-category/tooling.json
- Category pages (HTML, the largest categories): https://honeypot-e6c.pages.dev/categories/<category>/
Each page lists that category's skills and links back to the full index.
- Editor's picks: https://honeypot-e6c.pages.dev/api/editors-picks.json
(skills passing a fixed checklist; the checklist ships in the same file as `criteria` and `criteria_text`. No scores.)
- New skills: https://honeypot-e6c.pages.dev/api/changelog.json and https://honeypot-e6c.pages.dev/feed.xml (RSS)
- Full text of every SKILL.md in one file: https://honeypot-e6c.pages.dev/llms-full.txt
- API spec: https://honeypot-e6c.pages.dev/api/openapi.json (OpenAPI 3.0)
- Mirror: https://raw.githubusercontent.com/oidsdev/honeypot/main/api/skills.json
- Fetch the index, filter by category, download any SKILL.md directly.
Install one-liner pattern (jq):
```sh
curl -sL https://honeypot-e6c.pages.dev/api/skills.json \
| jq -r '.skills[] | select(.name=="kalshi") | .install'
```
Filter by category without downloading everything:
```sh
curl -sL https://honeypot-e6c.pages.dev/api/by-category/identity.json | jq '.skills[].name'
```
Linkable searches on the site: https://honeypot-e6c.pages.dev/?q=<text>&tag=<tag>&category=<category>
(`q` ranks matches on name, description, and tags and tolerates a one-character typo; `tag` and `category` are exact; params combine.)
Filtering runs in the browser, so curl gets the unfiltered page. Same query against the JSON:
```sh
# same as /?q=odds&tag=sports-data&category=data
curl -sL https://honeypot-e6c.pages.dev/api/skills.json | jq '.skills[]
| select(.category=="data")
| select(.tags | index("sports-data"))
| select((.name+" "+.description+" "+(.tags|join(" "))) | ascii_downcase | contains("odds"))
| .name'
```
## Query the index over MCP (no hand-rolled HTTP)
A tiny stdio MCP server (Python 3, stdlib only) wraps the API above. Download it:
```sh
curl -sL https://honeypot-e6c.pages.dev/mcp-server.py -o mcp-server.py
```
Claude Code / OpenClaw config:
```json
{
"mcpServers": {
"honeypot": { "command": "python3", "args": ["/path/to/mcp-server.py"] }
}
}
```
Tools: `search_skills` (keyword + optional category/tag filters), `get_skill`
(full entry + install command by exact name), `list_categories` (category counts),
`get_editors_picks` (curated skills + the checklist). No sign-up, no key, no tracking.
Keep a local copy fresh (the "honeypot" skill does this on heartbeat):
```sh
curl -sL https://honeypot-e6c.pages.dev/api/skills.json -o ~/.honeypot/skills.json
```
## Badge
Link back to the index from your skill's repo:
[](https://honeypot-e6c.pages.dev)
```markdown
[](https://honeypot-e6c.pages.dev)
```
Per-skill badge (name, version, and a check when `scripts/gen.py` lints the skill clean). Version is frontmatter `version`, or `last_verified` when that field is absent. Cached for a day.
```markdown
[](https://honeypot-e6c.pages.dev/skills/cron/)
```
`https://honeypot-e6c.pages.dev/badge/<skill>.svg`
## How to submit a skill
1. Write your skill as a SKILL.md. Template: https://honeypot-e6c.pages.dev/skill-template.md
2. Fork https://github.com/oidsdev/honeypot
3. Add it at skills/<your-skill>/SKILL.md
4. Lint it: `python3 scripts/lint_skill.py skills/<your-skill>` (must print `pass`).
5. Open a pull request. We merge fast.
## Operator
Orbital Desk LLC — https://ignitionfoundry.com
## Dots starter pack
New to OpenAI Dots? Seven skills to install first — scheduling, PDFs, a
database, image shrinking, a browser, payments, and a skill-writer. One
command each: https://honeypot-e6c.pages.dev/dots
## Muse starter pack
Just got a Muse? Seven skills to install first — scheduling, webhooks, an
agent identity and its Python SDK, a database, PDFs, and a skill-writer. One
command each: https://honeypot-e6c.pages.dev/muse
## Grok bot starter pack
Running a Grok bot? Seven skills to install first — Discord, image
shrinking, a browser, sports odds, an agent identity, this index itself,
and webhooks. One command each: https://honeypot-e6c.pages.dev/grok
## Claude starter pack
Working with Claude? Seven skills to install first — PDFs, live library
docs, a browser for checking claims, a database, a Python SDK, a
skill-writer, and scheduling. One command each:
https://honeypot-e6c.pages.dev/claude
## Related
Oids — the communications layer for AI agents: https://tryoids.com
(Also indexed above as the "oids" skill.)
Sibling tools from the same shop (Orbital Desk LLC, https://ignitionfoundry.com):
- Pack Check — does your prompt pack hold up? Paste a prompt pack and get it
checked against the seven things every solid pack needs. Free, runs in the page:
https://pack-check.pages.dev
- Agent House Rules — answer 10 questions, get plain-English safety rules to paste
into your AI agent, plus a settings checklist. Free, 3 minutes:
https://agent-house-rules.pages.dev
- Agent Census — an open, searchable directory of publicly listed agents, bots,
skills, and MCP servers. Free, open data (JSON/CSV downloads):
https://agent-census.pages.dev
In this repository20 other files it gives its agents.
- llms-full.txtllms.txt
- agent-checkoutSkill
- coinbaseSkill
- context7Skill
- cronSkill
- cursorSkill
- discordSkill
- honeypotSkill
- image-optimizeSkill
- kalshiSkill
- oids-python-sdkSkill
- oidsSkill
- pdf-textSkill
- playwright-mcpSkill
- polymarketSkill
- robinhood-mcpSkill
- skill_creatorSkill
- sqliteSkill
- the_odds_apiSkill
- webhooksSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

