Writes and edits the rules in an instruction file, whether that is `AGENTS.md`, the `CLAUDE.md` symlinked to it, or a preset's `AGENTS.partial.md`. Covers the bullet a rule lands as, how its reason and its exception attach, which rules earn a place, where a new one goes, and when a rule replaces an existing bullet instead of joining it. Invoked explicitly as `/agents-md <the rule or the change>`, and applies on its own whenever the deliverable is a rule in one of those files. Not for a skill…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# agents-md
**A rule that lands as one bullet in the section that already owns its domain, carries its own reason and its own exception, and either replaces an overlapping bullet or is not added at all.**
An instruction file is `AGENTS.md`, the `CLAUDE.md` symlinked to it, or a preset's `AGENTS.partial.md`. The rules below hold for all three. Which file a change belongs in is settled by the scope the rule holds for, which the request names rather than this skill.
## What a rule looks like
Every rule is one bullet under an H2 heading, and the file is a flat list of those bullets rather than prose. A rule inside a paragraph is read as background and a rule on its own line is read as a rule.
A rule addresses the agent directly, usually in the imperative. "Never use em-dashes" and "Cut rationale unless load bearing" are the form that gets followed. This is the one place an instruction file inverts a skill body, which states a property of a finished artifact instead, because an instruction file governs behavior rather than describing a thing.
A bullet runs one to three sentences. The first is the rule and the rest carry its reason, the failure it prevents, or its exception.
…
The line at the top of this page pins the label git:20260910.774b1b6, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
GET https://markdownregistry.com/api/v1/artifacts/art_6mxpx5vds6ly3uay
GET https://markdownregistry.com/api/v1/resolve?ref=tanq16/claudex/agents-md
GET https://markdownregistry.com/api/v1/blob/f1a7388b8f0ce8aa524ec7e2ae4967395d69b75a72ec4831013e1999423f913c
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
tanq16/claudex · internal/embedded/default-skills/skill-creator/SKILL.md · User-invoked skill author. Writes or edits a skill (the frontmatter, the body, and any reference files) in the Agent…
tanq16/claudex · internal/embedded/default-skills/write-document/SKILL.md · Settles what a standalone document should be before it is written: whether a file is the right output at all, who…
tanq16/claudex · internal/embedded/presets/private/skills/chrome-extension/SKILL.md · Chrome extension structure, Manifest V3, the popup, content scripts, service workers, and storage. Use when creating an…
tanq16/claudex · internal/embedded/presets/private/skills/develop/SKILL.md · The entry point for any coding work in a project that has these skills installed - implementing a feature, changing or…
tanq16/claudex · internal/embedded/presets/private/skills/dockerize/SKILL.md · Containerizing any project - multi-stage builds, choosing a base image, packaging the runtime, clearing package-manager…
tanq16/claudex · internal/embedded/presets/private/skills/github-release-workflow/SKILL.md · The GitHub Actions release workflow and the semver commit-marker convention. Use when creating or changing…
tanq16/claudex · internal/embedded/presets/private/skills/go-cli-commands/SKILL.md · The surface a Go CLI command exposes - the baseline every tool starts at, the two purchases on top of it, flag…
tanq16/claudex · internal/embedded/presets/private/skills/go-cli-output/SKILL.md · The utils printer for Go CLI tools - the two output tiers behind --debug, the Print* API, table rendering, terminal…
tanq16/claudex · internal/embedded/presets/private/skills/go-cli-progress/SKILL.md · The progress surface for Go CLI tools - the two-line live block, the in-place meter carrying rate and ETA, how it…
tanq16/claudex · internal/embedded/presets/private/skills/go-cli-prompts/SKILL.md · Interactive input for Go CLI tools - when a prompt is the right channel at all, how a value resolves from a flag then…
tanq16/claudex · internal/embedded/presets/private/skills/go-cli-structure/SKILL.md · The shape of a Go CLI's command tree - main.go, the root command for each project type, Execute and the exit path, a…
dimetron/pi-go · internal/extension/bundled_skills/agents-md/SKILL.md · Generate AGENTS.md files for Go, Rust, TypeScript, and Java projects. Use this skill whenever the user asks to create…
mblode/agent-skills · skills/agents-md/SKILL.md · Audits and edits agent instruction files, verifies repository commands, and migrates repositories to AGENTS.md as the…
outlinedriven/outline-driven-development · .devin/skills/agents-md/SKILL.md · Use when a repo needs agent setup, AGENTS.md added or made lean, CLAUDE.md audited, or agent instructions scored or…
outlinedriven/odin-claude-plugin · plugins/odin-agent/skills/agents-md/SKILL.md · Use when a repo needs agent setup, AGENTS.md added or made lean, CLAUDE.md audited, or agent instructions scored or…
wpfleger96/ai-agent-rules · src/ai_rules/config/skills/agents-md/SKILL.md · Create or update AGENTS.md with repo-specific patterns, conventions, commands, and gotchas for LLM coding agents. Use…
