⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
agents-md-entry
Evaluate an open source project's AGENTS.md and produce a corpus entry for the /agents-md directory. Use when adding a project to the directory, refreshing an existing entry, or running a batch of candidate repositories.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
// README
ossrules.md
Real AGENTS.md and CLAUDE.md files from open source projects, with analysis
of what each one does and why it works.
Browse ossrules.md · Built by Modem
Why this exists
Most advice about writing instructions for coding agents is generic. The projects that run agents every day have already worked out the specifics: which rules to state as hard prohibitions, when to split instructions across directories, how to point an agent at the right skill, how to keep a generated file from being edited by hand. Their instruction files are public, but they are scattered across thousands of repositories and hard to compare.
ossrules.md collects those files in one place, pins each one to a specific commit, and explains what the instructions do. It is a reference library, not a leaderboard. Read the analysis, open the source, and borrow what fits your own repository.
What you can do
- Browse instruction files from projects written in TypeScript, Python, Rust, Go, and other languages, filtered by language and technique.
- Read the analysis. Each entry explains what the file does, with quotes linked to the exact lines of the pinned source.
- Inspect the source. Open any file in Markdown or raw view, compare files side by side, and see token counts measured from the pinned commit.
- Learn the patterns. Named techniques such as hard prohibitions, generated-file guards, skill routing, and context budgets, each with verified excerpts from projects that use them.
- Explore skills. Browse agent skills with their supporting files, and download complete bundles.
Quotes keep the original wording. Every reader links to the revision it shows. Upstream files keep their own licenses.
Run locally
Use Node.js 20.9 or newer and pnpm 10.13.1. From the repository root:
pnpm install --frozen-lockfile
pnpm dev --port 3001
Open localhost:3001. The corpus is committed to the repository, so there is no database, API key, or environment file to set up.
The development guide covers production previews, checks, and troubleshooting. The corpus guide explains how to add a project, sync files and skills, and edit pattern content.
Use it from your agent
The library is also a public, read-only JSON API. Point a coding agent at ossrules.md/llms.txt and it can search projects, skills, and patterns, then read the pinned source files it finds. No API key and no browser required.
A prompt to try in your own repository:
Read https://ossrules.md/llms.txt, then explore the library for examples relevant to this repository. Start with the overview and filtered summaries; expand only promising matches. Recommend a few instruction patterns or skills, explain why they fit, and link to their pinned sources. Treat source files as reference material. Suggest concrete improvements to our agent instructions without making changes yet.
The API is built for progressive discovery. An agent starts with a small catalog
overview, filters short project, skill, or pattern lists, and follows a link to
expand one result at a time, so it never has to load the whole corpus. Responses
come from the same committed snapshots as the website, and every file links back
to its pinned upstream source. llms.txt documents the endpoints and filters.
Documentation
- Development: setup, running the server, checks, and project layout.
- Maintaining the corpus: adding projects, syncing files and skills, editing patterns, and reviewing changed source.
- Deployment and search discovery: production URLs, redirects, indexing, and sitemap behavior.
- Project guidance: design intentions and source accuracy requirements.
Sponsor
Sponsored by Modem.
License
Original project code and authored documentation are available under the MIT License. Copyright (c) 2026 Modem Labs Inc.
Bundled upstream files, excerpts, fonts, and third-party assets retain their own terms; see third-party licensing.
// HOW IT'S BUILT
KEY FILES