kayforkind/skill-slice/agents-md
Claude Code and agent skill by kayforkind/skill-slice
- community
Checking that the SKILL.md file is still in the repository.
Creates and audits lean AGENTS.md / CLAUDE.md agent instruction files (target under 60 lines, hard max 100) from the real toolchain, with a CLAUDE.md symlink. Use when the user asks to create, update, or audit AGENTS.md or CLAUDE.md, or to keep agent docs from duplicating linters. Not for authoring SKILL.md skills (create-skill), Claude Code plugin skill packs (skill-development), or README/CONTRIBUTING prose.
Install
npx vskill@latest install kayforkind/skill-slice/agents-mdSKILL.md
The same SKILL.md was published earlier as kayforkind/skill-slice/agents-md.
Maintaining AGENTS.md
AGENTS.md is the canonical agent-facing documentation file. Keep it minimal—agents are capable and don't need hand-holding. Target under 60 lines; never exceed 100. Instruction-following quality degrades as document length increases.
When to Use
- The user asks to create, update, or audit
AGENTS.mdorCLAUDE.md. - The project needs concise, high-signal agent instructions derived from the actual toolchain and repo layout.
- Existing agent documentation is too long, duplicated, or drifting away from real project conventions.
Prerequisites
- Project root must be accessible (this is where
AGENTS.mdlives). - On Windows hosts (primary), use PowerShell for symlink creation and file operations.
Procedure
1. File Setup
- Create
AGENTS.mdat project root. - Create a symlink so
CLAUDE.mdresolves to the same content.
Windows (PowerShell):
Linux / macOS:
On Windows, creating symlinks may require Developer Mode enabled or an elevated PowerShell session.
2. Analyze the Project Before Writing
Before writing anything, inspect the repository to understand what belongs in the file:
- Package manager — Check for lock files:
pnpm-lock.yaml,yarn.lock,package-lock.json,uv.lock,poetry.lock. - Linter/formatter configs — Look for
.eslintrc,biome.json,ruff.toml,.prettierrc, etc. Do not duplicate these rules in AGENTS.md. - CI/build commands — Check
Makefile,package.jsonscripts, CI configs for canonical commands. - Monorepo indicators — Check for
pnpm-workspace.yaml,nx.json, Cargo workspace, or subdirectorypackage.jsonfiles. - Existing conventions — Check for
CONTRIBUTING.md,docs/, orREADMEpatterns.
3. Writing Rules
- Headers + bullets — No paragraphs.
- Code blocks — For commands and templates.
- Reference, don't embed — Point to existing docs: "See
CONTRIBUTING.mdfor setup" or "Follow patterns insrc/api/routes/". - No filler — No intros, conclusions, or pleasantries.
- Trust capabilities — Omit obvious context.
- Prefer file-scoped commands — Per-file test/lint/typecheck commands over project-wide builds.
- Don't duplicate linters — Code style lives in linter configs, not AGENTS.md.
4. Required Sections
Package Manager
Which tool and key commands only:
File-Scoped Commands
Per-file commands are faster and cheaper than full project builds. Always include when available:
Commit Attribution
Always include this section. Agents should use their own identity:
Co-Authored-By: (the agent model's name and attribution byline)
Key Conventions
Project-specific patterns agents must follow. Keep brief.
5. Optional Sections
Add only if truly needed:
- API route patterns (show template, not explanation)
- CLI commands (table format)
- File naming conventions
- Project structure hints (point to critical files, flag legacy code to avoid)
- Monorepo overrides (subdirectory
AGENTS.mdfiles override root)
Pitfalls
- Length creep — Instruction-following quality degrades as document length increases. Target under 60 lines; never exceed 100.
- Duplicating linter rules — Never copy
.eslintrc,biome.json, orruff.tomlrules into AGENTS.md. Reference the config instead. - Embedding instead of referencing — Don't paste full docs; point to them ("See
CONTRIBUTING.md"). - Project-wide builds over file-scoped commands — Always prefer per-file test/lint/typecheck commands when available.
- Listing installed skills or plugins — Agents discover these automatically; don't list them.
- Windows symlink failures — On Windows, symlink creation may fail without Developer Mode or elevation. If it fails, create
CLAUDE.mdas a copy or useCopy-Itemas a fallback. - Anti-patterns to omit:
- "Welcome to..." or "This document explains..."
- "You should..." or "Remember to..."
- Full project-wide build commands when file-scoped alternatives exist
- Obvious instructions ("run tests", "write clean code")
- Explanations of why (just say what)
- Long prose paragraphs
Verification
- Line count check — Ensure the file is under 60 lines (hard max 100):
``powershell (Get-Content AGENTS.md).Count ` `bash wc -l AGENTS.md ``
- Symlink check — Confirm
CLAUDE.mdresolves toAGENTS.md:
``powershell Get-Item CLAUDE.md | Select-Object Target ` `bash ls -l CLAUDE.md ``
- No linter duplication — Grep for common linter rule keywords that should not appear:
``powershell Select-String -Path AGENTS.md -Pattern "semi","no-unused","max-len","indent" | Select-Object LineNumber,Line `` If matches are found (outside of legitimate command examples), remove them.
- Required sections present — Confirm all four required sections exist:
``powershell Select-String -Path AGENTS.md -Pattern "Package Manager","File-Scoped Commands","Commit Attribution","Key Conventions" ``
Example Structure
Co-Authored-By: (the agent model's name and attribution byline)
Limitations
- Use this skill only when the task clearly matches the scope described above.
- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/agents-md/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/kayforkind/skill-slice/agents-md)