llms-txt-writer
llms.txt3 starsChanged yesterdayMITView on GitHub
A Claude Code skill (Python, MIT, v0.2.0) that writes documents optimized to be cited by AI search engines (ChatGPT, Perplexity, Gemini). Follows Answer.AI's llms.txt standard and checks each draft with GEO-SFE 3-layer static analysis backed by Victorino LLC, The Digital Bloom, Position Digital, Omniscient Digital, and arXiv:2603.29979. Status: active; synced one way from the author's Claude Code harness via scripts/sync-from-local.sh.
In this repository2 other files it gives its agents.
- llms-full.txtllms.txt
- llms-txt-writerSkill
The file
Files
# llms-txt-writer > A Claude Code skill (Python, MIT, v0.2.0) that writes documents optimized to be cited by AI search engines (ChatGPT, Perplexity, Gemini). Follows Answer.AI's `llms.txt` standard and checks each draft with GEO-SFE 3-layer static analysis backed by Victorino LLC, The Digital Bloom, Position Digital, Omniscient Digital, and arXiv:2603.29979. Status: active; synced one way from the author's Claude Code harness via `scripts/sync-from-local.sh`. ## Core documentation - [llms-full.txt](llms-full.txt): self-contained Q&A reference for AI agents (Project Facts, Prior Research, 9 question-form H2 sections) - [README.md](README.md): human-facing narrative (install, how it works, GEO metrics, examples) - [SKILL.md](skills/llms-txt-writer/SKILL.md): full skill specification — when to use, post-script interpretation, ski-ramp optimization techniques 1 / 2 / 3, anti-patterns - [CHANGELOG.md](CHANGELOG.md): release notes (v0.2.0 initial release) - [scripts/sync-from-local.sh](scripts/sync-from-local.sh): one-way export of the skill from the author's harness into this repo ## Architecture - [scripts/geo_check.py](skills/llms-txt-writer/scripts/geo_check.py): bilingual (EN / JA) static analyzer — 5 GEO metrics in one pass, text or `--json` output; imports only the Python standard library (Japanese handled with regular expressions) - [pyproject.toml](skills/llms-txt-writer/pyproject.toml): Python 3.11+ with `uv`; no runtime dependencies (pytest for the tests) - [tests/test_geo_check.py](skills/llms-txt-writer/tests/test_geo_check.py): 48 passing unit + integration tests - [fixtures/](skills/llms-txt-writer/fixtures/): `sample_en.md`, `sample_ja.md` — minimal worked examples ## Standard reference - [llmstxt.org](https://llmstxt.org/): Answer.AI's `llms.txt` proposal (Jeremy Howard, 2024), which defines `llms.txt`; the `llms-full.txt` companion this skill also writes is a community convention (as of 2026-08-23) - [arXiv:2603.29979](https://arxiv.org/abs/2603.29979): GEO-SFE framework paper — macro / meso / micro analysis layers ## Sibling skills - [readme-writer](https://github.com/shimo4228/readme-writer): the human-facing counterpart — rewrites or reviews a README so first-time visitors can tell what the project is; a fresh judge checks the page against the code - [jsonld-knowledge-graph](https://github.com/shimo4228/jsonld-knowledge-graph): designs and ships a JSON-LD knowledge graph (`graph.jsonld`) next to `llms.txt` - [claude-skill-writing-ecosystem](https://github.com/shimo4228/claude-skill-writing-ecosystem): orchestrator for human-facing articles and essays - [Authorship Strategy](https://github.com/shimo4228/authorship-strategy): the parent practice line (concept DOI 10.5281/zenodo.20263316); its ADR-0006 asks for both a prose entry point (`llms.txt`) and a concept graph (`graph.jsonld`) - [shimo4228](https://github.com/shimo4228/shimo4228): the author's hub — the practice lines, their DOIs, and the tools
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.

