Infraestrutura de Agentes IA Org-wide para GitHub
petry-projects/.github-private
Agentes IA org-wide + automação de PR reviews + orchestração agentic
Ver no GitHubRepositório central que fornece agentes customizados do Copilot, skills do Claude Code e orquestração de workflows agentic para toda a organização petry-projects. Inclui automação de revisão de PRs em cascata, auditoria de conformidade e geração de ideias de features, tudo invocável via GitHub.com, VS Code e CLI.
Instalação
npx convez-ai petry-projects/.github-private1 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
Casos de uso
- Revisar PRs automaticamente em cascata (triage → análise profunda → auditoria de segurança)
- Auditar repositórios contra padrões organizacionais via agente compliance
- Gerar e priorizar ideias de features com IA generativa
- Orquestrar workflows de desenvolvimento com frameworks de agentes (BMAD, spec-kit, GSD)
Veredito

Solução enterprise de automação agentic bem-estruturada para governança de código e workflows em escala org. Requer expertise em GitHub Actions, prompt engineering e orchestração de agentes, mas oferece ROI alto em revisões automatizadas e padronização organizacional.
README
A tradução para o português deste README ainda está na fila.
.github-private
Org-wide Copilot custom agents, Claude Code skills, and agentic workflow infrastructure for petry-projects.
What This Repo Does
This is the .github-private convention repo for the petry-projects org. Agent profiles in /agents/ are available org-wide — invocable from GitHub.com, VS Code, JetBrains, and Copilot CLI.
It also contains the scheduled PR review automation (workflows + scripts + prompts) that runs hourly across the org.
Structure
agents/ # Copilot custom agent profiles (org-wide)
pr-reviewer.md # Multi-tier cascading PR review
feature-ideator.md # Feature idea generation
compliance-auditor.md # Org standards compliance checking
prompts/ # Prompt libraries used by workflows
scripts/ # Shell orchestration for GitHub Actions
frameworks/ # Installed agentic frameworks (git subtree)
bmad-method/ # BMAD-METHOD: multi-agent development lifecycle
spec-kit/ # spec-kit: spec → plan → tasks pipeline
gsd/ # get-shit-done: context engineering & milestones
.github/workflows/ # Scheduled automation (PR review, health checks)
Agents
| Agent | Purpose | Invoke with |
|---|---|---|
pr-reviewer | Cascading PR review (triage → deep → security audit) | @pr-reviewer in any org PR |
feature-ideator | Generate and prioritize feature ideas | @feature-ideator in any org repo |
compliance-auditor | Audit repo against org standards | @compliance-auditor in any org repo |
PR Review Automation
The scheduled workflow reviews all open PRs across petry-projects hourly:
- Classifies risk (LOW/MEDIUM/HIGH) via cascading tiers
- Auto-approves LOW/MEDIUM risk PRs with passing CI
- Cross-engine adversarial review (Claude + Copilot rubber duck)
- Escalates HIGH risk or failing PRs to human review
Trigger manually: gh workflow run pr-review.yml --repo petry-projects/.github-private
Mention trigger: Comment @petry-review-bot on any org PR for immediate review.