nonna
CLAUDE.md0 starsChanged 46 days agoMITView on GitHub
Nonna is a portable .claude/ harness that keeps AI-assisted development disciplined, test-driven, review-gated, and safe by construction. Language- and domain-agnostic: drop it into any repository and the same loop applies. The operating rules are in .claude/rules/, which loads on every turn. Start with 00-core.md — the three principles, the loop, the ladder, the never-list, and routing. This file holds only what lives nowhere else. Senior-staff engineering rigor. Every change is planned, tested first, reviewed, and verified before it merges. We…
In this repository29 other files it gives its agents.
- hosts/AGENTS.mdAGENTS.md
- adrSkill
- api-designSkill
- auditSkill
- code-reviewSkill
- concurrency-performanceSkill
- coverageSkill
- debuggingSkill
- debugSkill
- fast-laneSkill
- fixSkill
- implementSkill
- intakeSkill
- leanSkill
- migration-safetySkill
- nonnaSkill
- observabilitySkill
- planSkill
- refactoringSkill
- releaseSkill
- reviewSkill
- rollbackSkill
- security-reviewSkill
- shipSkill
- supply-chainSkill
- syncSkill
- tddSkill
- tdd-workflowSkill
- testSkill
The file
Files
CLAUDE.md — Nonna
Nonna is a portable .claude/ harness that keeps AI-assisted development disciplined, test-driven,
review-gated, and safe by construction. Language- and domain-agnostic: drop it into any repository
and the same loop applies.
The operating rules are in .claude/rules/, which loads on every turn. Start
with 00-core.md — the three principles, the loop, the ladder, the never-list, and
routing. This file holds only what lives nowhere else.
Caliber bar
Senior-staff engineering rigor. Every change is planned, tested first, reviewed, and verified before it merges. We do not ship hope. "Works on my machine" is not done; green tests + passing review + synced docs is done.
Model-tier policy
Match model to task depth — never burn a deep-reasoning model on mechanical work.
- Opus — planning, architecture, code/security review, debugging hard failures, orchestration.
- Sonnet — the bulk of implementation and test-writing.
- Haiku — high-volume mechanical work: fan-out search, bulk edits, simple single-file changes.
Rules precedence
Project rules in .claude/rules/ override any global ~/.claude/rules/. When in doubt, choose the
option that preserves correctness, safety, and reproducibility over cleverness or speed.
Where things live
.claude/README.md maps the harness. Agent and skill descriptions are already in context —
do not re-read an index to find them. Nonna installs as a plugin
(/plugin marketplace add kapadias/nonna); a plugin install is not equivalent to a copy-in
install; both default to lite mode — see docs/INSTALL.md,
ADR 0007 and ADR 0011.
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.

