claude-md-architect
Skill0 starsChanged 4 months agoMITView on GitHub
Author, restructure, or prune CLAUDE.md files for any repository. Use whenever the user asks to create or improve a CLAUDE.md, complains that Claude ignores project conventions, mentions config for a team/enterprise rollout, or when starting work in a repo whose CLAUDE.md is missing, auto-generated, or over 150 lines. Also use when asked to add rules after Claude makes a repeated mistake.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- aimasteracc/harness-kittheir own version, first seen Jun 3, 2026
- trin-zenityx/claude-md-architecttheir own version, first seen May 27, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- trin-zenityx/claude-md-architect★ 2same name
- aimasteracc/harness-kit★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository11 other files it gives its agents.
- angular-enterpriseSkill
- angular-migrationSkill
- characterization-testsSkill
- code-archaeologySkill
- feature-flag-changeSkill
- frontend-review-checklistSkill
- rxjs-modernizeSkill
- session-hygieneSkill
- session-stateSkill
- strangler-migrationSkill
- tdd-loopSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Must contain, in this order: Stack (exact versions), Commands, Conventions, Never, Repo map (≤ 15 lines, only if no harness nav map output …
- Very specific: line 13
Never, Repo map (≤ 15 lines, only if no harness nav map output exists).
- Distinctive: line 14
Every rule ≤ 2 sentences, imperative voice, testable.
- Distinctive: line 38
Add the narrowest rule that prevents X, place it in Never or Conventions, note the date and reason as an HTML comment, and check whether an …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- owner: Muskan, added: 2026-07-16, model: claude-sonnet-5, review: quarterly -->CLAUDE.md Architect
Non-negotiable output constraints
- Repo-level CLAUDE.md ≤ 150 lines. If content exceeds this, push detail into directory-level CLAUDE.md files or docs/ files referenced with one line.
- Must contain, in this order: Stack (exact versions), Commands, Conventions,
Never, Repo map (≤ 15 lines, only if no
harness nav mapoutput exists). - Every rule ≤ 2 sentences, imperative voice, testable. "Write good code" is banned; "Components must not inject services other than facades" is the bar.
- Stack lines require exact versions: "Angular 17.3, NgRx 17 (signals store), TypeScript 5.4". Reject "Angular" bare.
Procedure
- Inventory: read existing CLAUDE.md at all levels (system, repo root, subdirs), package manifests, lint configs, CI files. List conflicts.
- Interview the repo, not the user: derive commands from package.json / Makefile / CI. Only ask the user for things code cannot answer (e.g., "which of these two patterns is the sanctioned one?").
- Draft with the section order above. The Never section comes from: past incidents, lint suppressions found in code, and any "repeated mistake" the user mentioned. Never section is the highest-value real estate; spend effort here.
- Prune pass: delete any rule that (a) duplicates a linter the repo runs, (b) restates framework defaults, (c) compensates for an old model limitation (flag these to the user instead of silently keeping).
- Layering: rules applying to one directory move to that directory's CLAUDE.md. Personal preferences go to CLAUDE.local.md (git-ignored).
- Output the file(s), then a 5-line changelog: added / removed / moved / flagged-as-rot / open questions.
Maintenance mode (when asked to "add a rule because Claude did X wrong")
Add the narrowest rule that prevents X, place it in Never or Conventions, note the date and reason as an HTML comment, and check whether an existing rule already covers it (if so, sharpen that rule instead of adding).
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.