claude-code-config
CLAUDE.md1 starsChanged 2 months agoMITView on GitHub
Source of truth for this machine's Claude Code setup. install.sh copies everything here into ~/.claude/; the installed copy is what actually runs. The global rules that apply while working in this repo are not in this file. They live in global/CLAUDE.md, which is already loaded from ~/.claude/CLAUDE.md. Keep this file to repo-specific notes only — anything duplicated from global/CLAUDE.md is loaded twice per session here. Edit the source under this repo, then run ./install.sh. Editing ~/.claude/ directly is overwritten on…
In this repository10 other files it gives its agents.
- global/CLAUDE.mdCLAUDE.md
- agent-reviewSkill
- context-budgetSkill
- exploreSkill
- pr-createSkill
- retrospectSkill
- security-scanSkill
- simplifySkill
- test-genSkill
- triangulateSkill
The file
Files
claude-code-config
Source of truth for this machine's Claude Code setup. install.sh copies
everything here into ~/.claude/; the installed copy is what actually runs.
The global rules that apply while working in this repo are not in this
file. They live in global/CLAUDE.md, which is already loaded from
~/.claude/CLAUDE.md. Keep this file to repo-specific notes only — anything
duplicated from global/CLAUDE.md is loaded twice per session here.
Editing config
Edit the source under this repo, then run ./install.sh. Editing
~/.claude/ directly is overwritten on the next install, and
block-sensitive-files.sh blocks writes to the installed hooks, settings,
and CLAUDE.md for that reason.
| Edit here | Lands at |
global/CLAUDE.md, global/RTK.md | ~/.claude/ |
settings.json, hooks/, skills/, rules/ | ~/.claude/ |
Context budget
rules/common/*.md declare no paths: frontmatter, so they are auto-injected
into every session (~1,200 tokens) — keep them short. The language overlays
under rules/{lang}/ declare paths: and stay out of context until a
matching file is edited. Adding paths: to a common/ file silently drops it
out of the always-loaded set; tests/install.bats asserts they have none.
Tests
bats tests/ # full suite
bats tests/install.bats # installer contractHook changes need a matching test — most hooks are security gates, and an unexercised gate reports PASS by never running.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

