Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/benmarte/autoimprove/claude-mdgit clone --depth 1 https://github.com/benmarte/autoimproveWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/benmarte/autoimprove/claude-md)<a href="https://agentmods.dev/instructions/benmarte/autoimprove/claude-md"><img src="https://agentmods.dev/badge/instructions/benmarte/autoimprove/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00414 | $0.00414 |
| Opus 5 | $0.00207 | $0.00207 |
| Sonnet 5 | $0.00083 | $0.00083 |
| Haiku 4.5 | $0.00041 | $0.00041 |
Grade A, and why
autoimprove CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
89% identical to codemunch CLAUDE.md — 8 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
What it actually says
Project Rules
Release Checklist
Before cutting a new release, ALL of the following must be completed. Do not create the GitHub release until every item is done.
1. Changelog
- Create or update
CHANGELOG.mdwith a new entry for the version - Use Keep a Changelog format: Added, Changed, Fixed, Removed
- Every commit since the last release must be accounted for
2. Version Bump
- Update
VERSIONfile to the new version number - Follow semver: breaking changes = major, new features = minor, bug fixes = patch
3. README
- Update README.md to reflect any new features, changed behavior, or removed functionality
- Update any version numbers referenced in examples (e.g., update check example)
- Add new commands/options to the Commands table if applicable
- Update the "Plugin structure" tree if files were added/removed
4. Pre-release Verification
git statusmust be clean (all changes committed)- Review
git log --oneline v{LAST}..HEADto confirm nothing was missed - Verify the plugin structure is valid (all skills have SKILL.md, hooks are executable)
5. Create Release
- Commit all of the above first
- Push to main
- Create GitHub release with
gh release create v{VERSION} - Release notes should summarize changes (not duplicate the full changelog)
- Include a link to the full changelog in the release notes
Files to Never Modify During Automated Runs
LICENSE.claude-plugin/marketplace.jsonautoimprove.png
Commit Style
- Use imperative mood: "Add feature" not "Added feature"
- Keep subject line under 72 characters
- Include
Co-Authored-By: Claude Opus 4.6 <[email protected]>when Claude writes the commit
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 43 lines · 414 tokens per session scan A 9c7559bcb243
autoimprove CLAUDE.md is an instructions file published in the GitHub repository benmarte/autoimprove (5 stars, last pushed 5mo ago), licensed MIT. It adds 414 tokens to every session, about $0.0021 per session on Opus 5. A static security scan graded it A with 0 findings. It is 89% identical to codemunch CLAUDE.md, differing in 8 lines, and is treated as a copy.
Other instructions, from other repositories
codemunch CLAUDE.md
Instructions for benmarte/codemunch, covering project rules, release checklist, 1. changelog, 2. version bump and 3. readme.
kitaru-skills CLAUDE.md
Instructions for zenml-io/kitaru-skills, covering claude.md, what this repository distributes, editing the skill, branching and releases and local claude code testing.
inquisitor AGENTS.md
AGENTS.md instructions for iamalisson/inquisitor, covering agents.md — inquisitor (always-on), don't classify — probe (depth is an output), delegate before you dig (router first), the gate (objective — only raises depth, never lowers) and confidence check (behavioural, before you act).
living-manual CLAUDE.md
Instructions for dougstanford/living-manual, covering living-manual, start of every session, how work lands — non-negotiable, user's manual (living-manual) and releasing.
octowiz CLAUDE.md
Instructions for raelli/octowiz, covering octowiz, releasing & deploying, commands, agent skills and issue tracker.
goal-cc CLAUDE.md
Instructions for bullish0x/goal-cc, covering project notes, development and release checklist.