How do I install the Write Agents Md skill?

Run npx skillmds@latest add tomevault-io/write-agents-md in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Write Agents Md safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Write Agents Md?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Write Agents Md free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Write Agents Md?

tomevault-io (@tomevault-io) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Write Agents Md by tomevault-io — Agent Skill

Name: Write Agents Md
Author: tomevault-io

Write Agents Md

Use when Codex needs to create, rewrite, audit, slim down, or install AGENTS.md/CLAUDE.md agent instruction files for a repository, especially when avoiding /init-style generated context, reducing context debt, adding repo-specific pitfalls, or setting nested AGENTS.md guidance.

tomevault-io f125b73 2 files · 3.8 KB Updated Sep 17, 2026

File contents

Write AGENTS.md

Core Principle

Treat AGENTS.md as a compact map of project-specific traps and verification rules, not as a README for agents.

Only include information that changes agent behavior and is hard to infer from the repo itself.

Workflow

Identify the target scope.
- Root repo? Monorepo package? Existing CLAUDE.md mirror?
- If scope is unclear, ask before editing.
Read current evidence.
- Existing AGENTS.md, CLAUDE.md, README.md
- Package scripts, test config, CI config, lint/typecheck commands
- Nearby files only as needed to verify non-obvious conventions
Filter every candidate line.
- Keep: exact commands, required flags, forbidden substitutions, repeated agent mistakes, deployment/runtime traps, security/performance/UX guardrails, PR/commit checks.
- Remove: folder tours, stack lists, README duplicates, broad architecture essays, obvious style rules, team background.
Prefer nested files for large workspaces.
- Root AGENTS.md: routing and global rules only.
- Subdirectory AGENTS.md: package-specific traps and commands.
- The closest file should carry the most specific instruction.
Write in Korean by default unless the repo clearly uses English docs.
- Use bullets and short sections.
- Keep the file small. Aim for under 80 lines unless the repo genuinely needs more.
- Each changed line must trace to discovered repo evidence or user-provided constraints.
Verify.
- Re-read the final file.
- Check that commands are executable-looking and not invented.
- Mention any assumptions or missing commands instead of guessing.

Starter Install Command

To install a lean starter AGENTS.md in the current project:

agents-md-init

Useful variants:

agents-md-init /path/to/project
agents-md-init --print
agents-md-init --force

The command writes a cautious starter template only. After installing it, inspect the project and replace placeholders with evidence-backed rules.

Skill Install Command

To install this skill itself:

agents-md-skill install

Useful variants:

agents-md-skill install --scope user
agents-md-skill install --scope project
agents-md-skill install --agent claude --scope user
agents-md-skill install --agent gemini --scope project
agents-md-skill install --agent all --scope user
agents-md-skill install /path/to/project --agent codex --scope project

Default behavior is --scope user --agent codex. Use --force to overwrite an existing installed copy.

Reference

Read references/principles.md when you need the inclusion/exclusion checklist, anti-patterns, or examples.

Source: sdhong0609/agents-md-skill — distributed by TomeVault.

---name: write-agents-mddescription: Use when Codex needs to create, rewrite, audit, slim down, or install AGENTS.md/CLAUDE.md agent instruction files for a repository, especially when avoiding /init-style generated context, reducing context debt, adding repo-specific pitfalls, or setting nested AGENTS.md guidance.---# Write AGENTS.md## Core PrincipleTreat `AGENTS.md` as a compact map of project-specific traps and verification rules, not as a README for agents.Only include information that changes agent behavior and is hard to infer from the repo itself.## Workflow1. Identify the target scope. - Root repo? Monorepo package? Existing `CLAUDE.md` mirror? - If scope is unclear, ask before editing.2. Read current evidence. - Existing `AGENTS.md`, `CLAUDE.md`, `README.md` - Package scripts, test config, CI config, lint/typecheck commands - Nearby files only as needed to verify non-obvious conventions3. Filter every candidate line. - Keep: exact commands, required flags, forbidden substitutions, repeated agent mistakes, deployment/runtime traps, security/performance/UX guardrails, PR/commit checks. - Remove: folder tours, stack lists, README duplicates, broad architecture essays, obvious style rules, team background.4. Prefer nested files for large workspaces. - Root `AGENTS.md`: routing and global rules only. - Subdirectory `AGENTS.md`: package-specific traps and commands. - The closest file should carry the most specific instruction.5. Write in Korean by default unless the repo clearly uses English docs. - Use bullets and short sections. - Keep the file small. Aim for under 80 lines unless the repo genuinely needs more. - Each changed line must trace to discovered repo evidence or user-provided constraints.6. Verify. - Re-read the final file. - Check that commands are executable-looking and not invented. - Mention any assumptions or missing commands instead of guessing.## Starter Install CommandTo install a lean starter `AGENTS.md` in the current project:```bashagents-md-init```Useful variants:```bashagents-md-init /path/to/projectagents-md-init --printagents-md-init --force```The command writes a cautious starter template only. After installing it, inspect the project and replace placeholders with evidence-backed rules.## Skill Install CommandTo install this skill itself:```bashagents-md-skill install```Useful variants:```bashagents-md-skill install --scope useragents-md-skill install --scope projectagents-md-skill install --agent claude --scope useragents-md-skill install --agent gemini --scope projectagents-md-skill install --agent all --scope useragents-md-skill install /path/to/project --agent codex --scope project```Default behavior is `--scope user --agent codex`. Use `--force` to overwrite an existing installed copy.## ReferenceRead `references/principles.md` when you need the inclusion/exclusion checklist, anti-patterns, or examples.---> Source: [sdhong0609/agents-md-skill](https://github.com/sdhong0609/agents-md-skill) — distributed by [TomeVault](https://tomevault.io).

Frequently asked questions