batch-issues
Batch-process open GitHub issues with parallel agents. Analyzes conflicts, dispatches workers, creates PRs. Usage: /batch-issues [issue numbers] [--dry-run] [--max-parallel N]
Informations de source
- Dépôt
- future-architect/uzomuzo-oss
- Dernière activité de la source
- 7 avril 2026 à 22:15
- Langue détectée de SKILL.md
- anglais
- Étoiles
- 34
- Forks
- 5
Options d'installation
Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.
Vérifiez les fichiers source
Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.
Affichage de SKILL.md
SKILL.md
Instructions source · Aperçu en lecture seule- name
- batch-issues
- description
- Batch-process open GitHub issues with parallel agents. Analyzes conflicts, dispatches workers, creates PRs. Usage: /batch-issues [issue numbers] [--dry-run] [--max-parallel N]
- argument-hint
- Options: - No args: auto-select and process open issues - Issue numbers: '/batch-issues 168 171 187' to process specific issues - --dry-run: show dispatch plan without launching agents - --max-parallel N: limit parallel agents (default 4)
# /batch-issues -- Parallel Issue Dispatcher
> **Full specification**: See `.github/prompts/batch-issues.prompt.md` for the complete
> procedure, conflict analysis, agent workflow, and safety rules.
## Quick Reference
- **Auto-select**: `/batch-issues` -- pick open issues, analyze conflicts, dispatch
- **Specific issues**: `/batch-issues 168 171 187` -- process these issues
- **Dry run**: `/batch-issues --dry-run` -- show plan only
- **Limit parallelism**: `/batch-issues --max-parallel 2`
- Phase 1: Issue selection & filtering
- Phase 2: Conflict analysis (shared file detection)
- Phase 3: Size/risk assessment (S/M/L classification)
- Phase 4: Dispatch plan (present to user)
- Phase 5: Agent dispatch (worktree-isolated, parallel across groups, sequential within)
- Phase 6: Summary report with PR links
- Phase 6.5: Orchestrator review — run `/review-until-clean` on each PR (agents can't launch independent reviewers)
- Phase 7: Evidence verification — check/fix Before/After test output in every PR body
- Each agent: reproduce (MANDATORY) -> architect -> implement -> self-check -> verify after (MANDATORY) -> before/after PR (template enforced)
- Independent review runs at orchestrator level (Phase 6.5), NOT inside agents
Voir sur GitHub