convez.ai
SkillDevOps & InfraestruturaIntermediário
leonardocandiani

Keepwright

leonardocandiani/keepwright

Arquitetura e qualidade automática para qualquer projeto Git

Ver no GitHub
BomPluginScore · de 0 a 100

Skill Claude Code que implanta e mantém arquitetura de engenharia de alta qualidade em repositórios Git. Configura constituição, regras estruturadas, GitHub Actions com auto-review via Claude, validadores portáveis e hooks — detectando a stack automaticamente e adaptando a configuração.

Instalação

$npx convez-ai leonardocandiani/keepwright

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
1
Acompanhando
0
Saúde
85%
Último commit
há 1 mês
Versão
v2.3.0
Issues abertas
2
Contribuidores
3

Casos de uso

  • Estabelecer padrões de design e escrita em projetos existentes
  • Automatizar code review com CI/CD e agentes Claude
  • Auditar e manter conformidade arquitetural contínua
  • Detectar stack e instalar validators específicos por tecnologia

Veredito

Ferramenta robusta para equipes que precisam enforçar qualidade sistêmica em repositórios. Combina CI/CD, análise de padrões e validators em uma solução integrada — ideal para brownfield e projetos em crescimento, com overhead inicial compensado pela manutenção automática.

README

A tradução para o português deste README ainda está na fila.

keepwright

Set up and keep a high-quality engineering architecture in any git repo.

A Claude Code plugin that implants a constitution, structured rules, GitHub Actions (CI, AI PR review, @claude mention, safe auto-merge), portable validators, and git hooks — detecting your stack and adapting. After setup it keeps maintaining: it audits the repo and uses multi-agent workflows to derive your design and writing-voice patterns, then turns them into rules and validators.

Install

Run each /plugin command on its own — don't paste both at once.

1. Add the marketplace

/plugin marketplace add leonardocandiani/keepwright

2. Install the plugin

/plugin install keepwright

3. Reload to activate

/reload-plugins

Loads keepwright's commands, skills, and agents into the current session — no Claude Code restart needed.

Commands

CommandWhat it does
/keepwright:setupInteractive wizard. Detects the stack and installs the full architecture.
/keepwright:auditChecks integration coverage of an existing repo against the architecture.
/keepwright:reviewCompares repo state against the patterns derived from your code and docs.

Workflows

Multi-agent orchestration the commands run under the hood — each fans out parallel agents and synthesizes the result. You normally don't call these directly (the commands trigger them), but they're invocable on their own for advanced use:

WorkflowWhat it doesTriggered by
/keepwright:map-brownfieldParallel read-only analysis of a large repo, synthesized into a config enrichment.setup (large repos)
/keepwright:derive-patternsMines the repo's design + writing-voice patterns into rules and validator specs.setup, review
/keepwright:verify-setupAdversarially verifies a fresh setup in parallel: secrets, equalization, workflow YAML, validators, the P1–P5 hierarchy.audit --deep

Skills & agents

  • Skills — keepwright (the methodology behind the wizard) and pr-review (the review procedure the CI calls as /pr-review #N).
  • Agents — design-auditor and voice-auditor: read-only auditors that inspect the repo's design and writing-voice dimensions.

Three layers

  • Wizard (/keepwright:setup) — an interactive command that detects git, stack (Node/Deno/Python/etc), Claude config, and existing CI, then installs the constitution, rules, workflows, validators, and hooks. Destructive steps ask for explicit approval.
  • Engine — the deterministic part: portable validators and git hooks that run the same way on every machine and in CI. No model in the loop, no flaky output.
  • Workflows — multi-agent orchestration that audits an existing repo, derives its design and writing-voice patterns, and writes them back as rules and validators.

What it installs

  • Constitution — CLAUDE.md as an equalized index of the rules, with the always-loaded invariants inline.
  • Rules — .claude/rules/: invariants, pipeline equalization, the P1–P5 epistemic hierarchy, PR flow, lesson catalysis, parallel work streams, safe merge, empirical proof before merge, and issue triage.
  • GitHub Actions — ci.yml (type-check, lint, validators), pr-auto-review.yml (heuristic + Claude review over OAuth), claude-mention.yml (@claude on demand), pr-auto-merge.yml (auto-merge only for inert changes), issue-triage.yml (advisory labels via free GitHub Models), and a deploy template picked by stack.
  • Issue triage — new issues are classified by GitHub Models (free in Actions, no secret) and get advisory labels + a summary comment, deterministically. It never closes, assigns, or merges — least-privilege by construction, so a prompt injection in an issue body cannot reach code or secrets. Ships with issue templates and scripts/seed-labels.sh. Turn it off with "issues": { "triage": "off" }.
  • Validators — portable TypeScript checks: secret scanning, CLAUDE.md sync, epistemic-hierarchy gate, empirical-proof gate, webhook-active check.
  • Hooks — lefthook (pre-commit validators + type-check, conventional commit-msg, force-push guard on main) plus structure and TODO generators.

Auth

The AI workflows use CLAUDE_CODE_OAUTH_TOKEN. Run /install-github-app and pick the subscription/OAuth option — it wires the token for you. Without the secret, the AI jobs skip gracefully. An API key is a documented fallback, not recommended.

If you need to set the secret by hand, scripts/setup-oauth-secret.sh <owner>/<repo> reads the token from the macOS Keychain or CLAUDE_CODE_OAUTH_TOKEN, validates its shape, and sets it without mangling.

Double merge gate

Real auto-merge runs only for inert changes (docs, chronology, work-stream notes). Anything touching code, CI, rules, deploy, or config is human-gated: the AI prepares the PR, a human gives a one-line go. Detail in templates/rules/07-safe-merge.md.template.

License

MIT. See LICENSE and AUTHORS.md.

Ver o original no GitHub

Bom para

Relacionados