# Ag Corrigir Bugs

> Bugfix unificado com auto-routing. Triage (classificar bugs), Fix (1 bug + 5 gates), Batch (2-5 bugs em sprints), Parallel (6+ bugs em Agent Teams). Substitui ag-corrigir-bugs/24/25/26.

- **Type:** Skill
- **Install:** `agentstack add skill-andregusman-raiz-a-gusman-claude-ag-corrigir-bugs`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [andregusman-raiz](https://agentstack.voostack.com/s/andregusman-raiz)
- **Installs:** 0
- **Category:** [Communication](https://agentstack.voostack.com/c/communication)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [andregusman-raiz](https://github.com/andregusman-raiz)
- **Source:** https://github.com/andregusman-raiz/a-gusman-claude/tree/main/skills/ag-corrigir-bugs

## Install

```sh
agentstack add skill-andregusman-raiz-a-gusman-claude-ag-corrigir-bugs
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# ag-corrigir-bugs — Bugfix

Spawn the `ag-23-bugfix` agent to handle any bugfix workflow — from triage to parallel execution.

## Auto-Routing

The agent auto-selects the best mode based on input:

| Input | Mode | Behavior |
|-------|------|----------|
| Lista sem contexto / "triar" | `--triage` | Read-only classification + sprint plan |
| 1 bug claro | `--fix` | Single fix with 5 quality gates |
| 1 bug obscuro | ag-depurar-erro first | Then `--fix` |
| 2-5 bugs | `--batch` | Sequential sprints, incremental commits |
| 6+ independent bugs | `--parallel` | Agent Teams, exclusive ownership |
| 6+ bugs with file overlap | `--batch` | Sequential for safety |

## Invocation

Use the **Agent tool** with:
- `subagent_type`: `ag-23-bugfix`
- `mode`: `bypassPermissions`
- `run_in_background`: `true` (except `--fix` which runs foreground)
- `prompt`: Compose from template below + $ARGUMENTS

## Prompt Template

```
Projeto: [CWD or user-provided path]
Modo: [--triage|--fix|--batch|--parallel] (ou auto-detect)
Bugs: [lista de bugs — IDs, descricoes, path para diagnostico, ou inline]
Branch: [branch atual ou nome para criar]
Diagnostico: [path para bug_triage.md gerado por mode triage, se disponivel]

Executar bugfix no modo indicado (ou auto-detect baseado na quantidade de bugs).
Seguir todas as quality gates do modo selecionado.
```

## Important
- ALWAYS spawn as Agent subagent — do NOT execute inline
- After spawning, confirm to the user
- For `--triage`: agent is READ-ONLY, does NOT fix bugs
- For `--fix`: runs foreground (sequential quality gates, user needs result)
- For `--batch`: uses worktree isolation
- For `--parallel`: uses Agent Teams with exclusive file ownership
- Gates are blocking: if any gate fails, fix before proceeding

## Escalacao: Issues para Bugs Nao Resolvidos

### Mode --triage
Apos classificar bugs, spawnar ag-registrar-issue para cada bug que NAO sera fixado nesta sessao:
```
Agent({
  subagent_type: "ag-registrar-issue",
  name: "issue-registrar",
  model: "haiku",
  run_in_background: true,
  prompt: "Repo: [detectar]\nOrigem: ag-corrigir-bugs --triage\nSeveridade: [P0-P3 da triagem]\nTitulo: [bug title]\nContexto: [descricao + classificacao + sprint sugerido]\nArquivos: [arquivos afetados]\nLabels: bug, triage"
})
```

### Mode --fix / --batch / --parallel
Se um bug falha todas as tentativas de fix (gate nao passa):
- Registrar via ag-registrar-issue com contexto das tentativas falhadas
- Continuar com proximo bug (nao bloquear batch/parallel)

## Migration Note

This agent replaces:
- ag-B-25 (diagnosticar-bugs) → mode `--triage`
- ag-corrigir-bugs (fix-verificar) → mode `--fix`
- ag-corrigir-bugs-batch → mode `--batch`
- ag-B-24 (bugfix-paralelo) → mode `--parallel`

All functionality, anti-patterns, quality gates, and workflows from the original agents are preserved in the merged agent definition.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [andregusman-raiz](https://github.com/andregusman-raiz)
- **Source:** [andregusman-raiz/a-gusman-claude](https://github.com/andregusman-raiz/a-gusman-claude)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-andregusman-raiz-a-gusman-claude-ag-corrigir-bugs
- Seller: https://agentstack.voostack.com/s/andregusman-raiz
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.