agents-md-improver
Audit and improve AGENTS.md files for Codex project memory.
Skill by chonamdoo
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/chonamdoo-agents-md-improver
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: security, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/agents-md-improver
What's inside
- Skill name
agents-md-improver- Description
- Audit and improve AGENTS.md files for Codex project memory. Use when the user asks to check, audit, update, optimize, shrink, deduplicate, or fix AGENTS.md, project instructions, repo memory, context bloat, context optimization, or agent rule quality. Finds AGENTS.md files, scores them for commands/workflows, architecture clarity, non-obvious patterns, conciseness, currency, and actionability, reports issues first, then makes targeted updates only after the user approves.
- License
- MIT
- Instructions
- 100 lines in SKILL.md · view on GitHub
Files (5)
- SKILL.md
- openai.yaml · agents/
- quality-criteria.md · references/
- templates.md · references/
- update-guidelines.md · references/
SKILL.md (first lines)
# AGENTS.md Improver Audit and improve Codex `AGENTS.md` files so future sessions get concise, current, actionable project context. ## Workflow ### Phase 1: Discover Find relevant instruction files without scanning generated folders: ```bash rg --files -g 'AGENTS.md' -g '!node_modules' -g '!build' -g '!dist' -g '!coverage' -g '!vendor' -g '!.git' ``` Include `~/.codex/AGENTS.md` only when the task explicitly concerns global Codex behavior. Common locations: | Location | Purpose | | --- | --- | | `./AGENTS.md` | Primary project context checked into the repo | | nested `AGENTS.md` | Package, app, or module-specific rules | | `~/.codex/AGENTS.md` | User-wide Codex defaults | | `.Codex/rules/` | Larger imported rules or compressed docs when the project uses that pattern | ### Phase 2: Assess Quality Read `references/quality-criteria.md` when scoring or producing an audit report. Score each file on: | Criterion | Weight | | --- | --- | | Commands and workflows | 20 | | Architecture clarity | 20 | | Non-obvious patterns | 15 | | Conciseness | 15 | | Currency | 15 | | Actionability | 15 | Flag stale commands, broken paths, duplicated instructions, generic advice, outdated architecture, TODOs that are not useful to future agents, and long content that should move to `.Codex/rules/` only when the project already uses that import pattern. ### Phase 3: Report Before Editing Always show the quality report before making edits. Use this shape: ```markdown ## AGENTS.md Quality Report ### Summary - Files found: X - Average score: X/100 - Files needing update: X ### File-by-File Assessment #### 1. ./AGENTS.md **Score: XX/100 (Grade: X)** | Criterion | Score | Notes | | --- | --- | --- | | Commands/workflows | X/20 | ... | | Architecture clarity | X/20 | ... | | Non-obvious patterns | X/15 | ... | | Conciseness | X/15 | ... | | Currency | X/15 | ... | | Actionability | X/15 | ... | **Issues:** - ... **Recommended updates:** - ... ``` ### Phase 4: Propose Targeted Updates Read `references/update-guidelines.md` before proposing changes. Only add information that will help future Codex sessions: - commands or workflows discovered during analysis - gotchas or non-obvious patterns found in the repo - package relationships not obvious from filenames - testing or verification approaches with evidence of actual execution; label declaration-only checks as unexecuted - configuration quirks that would otherwise be rediscovered Avoid generic best practices, one-off fixes, verbose explanations, and statements that merely restate file or class names. Show a small diff for each proposed change and ask for approval before editing. ### Phase 5: Apply After approval, edit the smallest relevant `AGENTS.md` file. Preserve existing structure, keep project files under 200 lines when feasible, and move longer material into `.Codex/rules/` with an import if the project already uses that pattern. Use `references/templates.md` only when a file is missing major sections or needs a clean structure.
From chonamdoo/agent-flow, MIT.
More from chonamdoo/agent-flow
- 0.570score
android-guides
chonamdoo
Shared Android reference bundle for other android-* skills, covering architecture, Compose, data, Gradle, DI, concurrency, testing, and review checklists.
SkillAFreshness: AClaude Code +9 - 0.565score
architecture
chonamdoo
Approved local architecture for the shop app.
SkillAFreshness: AClaude Code +9 - 0.565score
android-clean-architecture
chonamdoo
Android/Kotlin platform adapter for `clean-architecture-core` that maps the core contract to Gradle modules, Kotlin packages, Hilt, Retrofit data sources, and…
SkillAFreshness: AClaude Code +9 - 0.565score
android-appshell-error-handling
chonamdoo
Defines Android app-shell common-error handling where feature ViewModels notify common errors and AppShell owns global UI, root navigation, Navigation3 back…
SkillAFreshness: AClaude Code +9 - 0.565score
team-flutter-conventions
chonamdoo
Shop team conventions for Flutter feature code under lib/features — repositories, list state notifiers, DTO decoding, and error handling.
SkillAFreshness: AClaude Code +9 - 0.565score
android-debugging
chonamdoo
Android diagnosis workflow for crashes, ANRs, incorrect UI state, coroutine or Flow bugs, data loading failures, and Compose rendering issues.
SkillAFreshness: AClaude Code +9 - 0.565score
android-clean-presentation-architecture
chonamdoo
Defines Android Clean Architecture presentation-layer guidance for Hilt DI, ViewModel, StateFlow uiState, one-shot UI events, and Compose screen wiring.
SkillAFreshness: AClaude Code +9 - 0.565score
android-code-review
chonamdoo
Android Kotlin and Jetpack Compose review checklist for changed code, covering architecture boundaries, MVI/state correctness, Compose stability, coroutine…
SkillAFreshness: AClaude Code +9 - 0.561score
architecture
chonamdoo
Orders app architecture contract.
SkillAFreshness: AClaude Code +9 - 0.561score
architecture
chonamdoo
Approved architecture contract for this Next.js app — one flat folder per feature, role-named files, T-prefixed types, and no Clean layering.
SkillAFreshness: AClaude Code +9 - 0.561score
architecture
chonamdoo
Shop app architecture contract.
SkillAFreshness: AClaude Code +9 - 0.559score
rn-io-conventions
chonamdoo
Team I/O convention for React Native feature code.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 21
- Last commit (this folder)
- Sep 12, 2026 (1 month ago)
- Last release
- Oct 21, 2026 (9 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/chonamdoo/agent-flow
- Website
- agent-flow-delta.vercel.app
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from chonamdoo
- 0.570score
android-guides
chonamdoo
Shared Android reference bundle for other android-* skills, covering architecture, Compose, data, Gradle, DI, concurrency, testing, and review checklists.
SkillAFreshness: AClaude Code +9 - 0.565score
android-clean-architecture
chonamdoo
Android/Kotlin platform adapter for `clean-architecture-core` that maps the core contract to Gradle modules, Kotlin packages, Hilt, Retrofit data sources, and…
SkillAFreshness: AClaude Code +9 - 0.565score
architecture
chonamdoo
Approved local architecture for the shop app.
SkillAFreshness: AClaude Code +9 - 0.565score
android-debugging
chonamdoo
Android diagnosis workflow for crashes, ANRs, incorrect UI state, coroutine or Flow bugs, data loading failures, and Compose rendering issues.
SkillAFreshness: AClaude Code +9 - 0.565score
team-flutter-conventions
chonamdoo
Shop team conventions for Flutter feature code under lib/features — repositories, list state notifiers, DTO decoding, and error handling.
SkillAFreshness: AClaude Code +9 - 0.565score
android-appshell-error-handling
chonamdoo
Defines Android app-shell common-error handling where feature ViewModels notify common errors and AppShell owns global UI, root navigation, Navigation3 back…
SkillAFreshness: AClaude Code +9
Cite this page
agents-md-improver skill. Agent Mart. https://agentmart.dev/skills/chonamdoo-agents-md-improver (last updated 2026-10-24)