NemesLaszloNemesLaszlo/claude-code-harness/CLAUDE.md

claude-code-harness

CLAUDE.md1 starsChanged 11 days agoMITView on GitHub

AGENTS.md (imported above) owns the six-beat workflow, the multi-agent rules and the repo-specific sections. This file holds only what Claude Code needs on top of it. Research-only agents cannot edit files (disallowedTools). Clean Code Architect, Backend Architect and Frontend Developer preload the code-structure skill; the other builders do not need the actions/service split.

1 other project uses a version of this.

Adopted by

Most recently taken up by

NemesLaszlo/Agent-Collectionword for word, first seen Nov 1, 2026

Other versions1 project uses a version of this.
Word for word · 1
NemesLaszlo/Agent-Collection★ 1100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
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 repository16 other files it gives its agents.
- AGENTS.mdAGENTS.md
- before-and-afterSkill
- code-structureSkill
- cross-reviewSkill
- evidence-driven-testingSkill
- exploreSkill
- fix-issuesSkill
- greploopSkill
- implementSkill
- investigateSkill
- new-featureSkill
- plan-featureSkill
- reviewSkill
- security-scanSkill
- shipSkill
- unslopSkill

The file

We split this CLAUDE.md into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Very specific:
Never start /implement, /ship or /greploop yourself.
line 11
Very specific:
Never build on main. If the branch is main, point to /new-feature or claude -w <task> before editing.
line 12
Very specific:
After editing this file or AGENTS.md, run /doctor prompt-audit to catch conflicts.
line 14

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md 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.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

CLAUDE.md

Claude Code Harness

@AGENTS.md

AGENTS.md (imported above) owns the six-beat workflow, the multi-agent rules and the repo-specific sections. This file holds only what Claude Code needs on top of it.

Rules

Query Context7 MCP before answering anything about a library, framework, SDK, API, CLI tool or cloud service, even without being asked.
Models: agents that plan, investigate, review or design run on fable. Agents that build and edit files run on opus. /review pins fable on the forked Clean Code Architect. /cross-review runs fable and opus side by side. Never downgrade an agent to sonnet or haiku without asking.
Never start /implement, /ship or /greploop yourself; they are user-invoked (disable-model-invocation: true) because they change the repo or push commits in a loop. Tell the user which one to type.
Never build on main. If the branch is main, point to /new-feature or claude -w <task> before editing.
Apply /unslop to every commit message, PR body, doc edit and comment you write.
After editing this file or AGENTS.md, run /doctor prompt-audit to catch conflicts.

Agents

Research-only agents cannot edit files (disallowedTools). Clean Code Architect, Backend Architect and Frontend Developer preload the code-structure skill; the other builders do not need the actions/service split.

Agent	File	Role	Model
Feature Planner	`.claude/agents/engineering/feature-planner.md`	Three-phase plan validation (read-only)	fable
Clean Code Architect	`.claude/agents/engineering/clean-code-architect.md`	Quality implementation, refactoring, DeepDive handoffs, .NET 10	opus (fable under `/review`)
DeepDive	`.claude/agents/engineering/deepdive.md`	Root cause analysis (read-only)	fable
Security Vulnerability Scanner	`.claude/agents/engineering/security-vulnerability-scanner.md`	Security audit (read-only)	fable
Backend Architect	`.claude/agents/engineering-additional/backend-architect.md`	APIs, data, scaling	opus
Frontend Developer	`.claude/agents/engineering-additional/frontend-developer.md`	Angular 20+, React, Vue, accessibility	opus
Mobile App Builder	`.claude/agents/engineering-additional/mobile-app-builder.md`	React Native, Flutter	opus
AI Engineer	`.claude/agents/engineering-additional/ai-engineer.md`	LLM features, Agent SDK, MCP, Claude Code harnesses	opus
DevOps Automator	`.claude/agents/engineering-additional/devops-automator.md`	CI/CD, infrastructure, PR checks	opus
Rapid Prototyper	`.claude/agents/engineering-additional/rapid-prototyper.md`	MVPs and spikes	opus
UI Designer	`.claude/agents/design/ui-designer.md`	Design systems, component specs	fable
UX Researcher	`.claude/agents/design/ux-researcher.md`	Research and synthesis (read-only)	fable

Which skill runs which agent

Beat	Skill	Agent
Isolate	`/new-feature`	main context
Plan	`/plan-feature`	forked Feature Planner
Plan	`/investigate`	forked DeepDive
Plan	`/explore`	forked Explore
Build	`/implement`	Backend Architect, Frontend Developer, DevOps Automator, Clean Code Architect per task
Review	`/cross-review`	two general-purpose reviewers, fable + opus, launched in parallel
Review	`/review`	forked Clean Code Architect, pinned to fable
Review	`/fix-issues`	Backend Architect, Frontend Developer, Clean Code Architect per file
Prove	`/evidence-driven-testing`, `/before-and-after`	main context
Ship	`/ship`, `/greploop`	main context
Ship	`/security-scan`	forked Security Vulnerability Scanner

Hand-offs between agents

Bug: DeepDive (/investigate) → Clean Code Architect, with the handoff as its brief.
Security: Security Scanner (/security-scan) → /fix-issues.
Design: UX Researcher → UI Designer → Frontend Developer, proven with /before-and-after.
Architecture: Backend Architect → Clean Code Architect.
Prototype: Rapid Prototyper → Backend Architect or Frontend Developer.
Unknown code: /investigate or /explore → /plan-feature → /implement.

Sizing the work

Size	Path
Small (one task, obvious scope)	`/new-feature`, build directly, `/review`, `/ship`
Medium or large (several tasks, data, auth, APIs)	`/new-feature` → `/plan-feature` → `/implement` → `/fix-issues` → `/evidence-driven-testing` → `/ship` → `/security-scan`
Exploratory (unknowns)	`/investigate` or `/explore` first, then the medium path

Skill details, arguments and requirements live in README.md and each skill's SKILL.md.

Download now (ZIP)

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

What did you use it for? e.g. Reviewing pull requests in a Rails app

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.

Check your setup, free