Repository navigation
guardrails: safe agents and workflow commands #5
Description
Parent epic: #1
Source brief:
docs/ai-guardrails/issues/004-safe-agents-and-commands.mddocs/ai-guardrails/migration/claude-code-skills-inventory.md
Problem
Raw built-in agents are too permissive for an internal product. The repo needs a safer default operating model for implementation, review, and release workflows.
Deliverables
- hardened default primary agent
- review-oriented subagent
- slash commands for
/implement,/review,/ship, and/handoff - explicit permission policy for dangerous shell patterns and write operations
Acceptance
- default agent is not an unrestricted build clone
- review workflow can run without edit access
- release workflow cannot bypass explicit gates
Notes
- Follow the thin-distribution approach from
docs/ai-guardrails/adr/001-thin-distribution-over-deep-fork.md - Preserve the philosophy imported from
claude-code-skillsepic feat(guardrails): Wave 8 — review fixes + remaining hooks + multi-model delegation #130: mechanism-first guardrails, fast feedback, pointer-based instructions, and runtime verifiability - Prefer OpenCode-native config/profile/plugin/command/CI surfaces over core patches
Dependencies
- epic: internal AI guardrails thin distribution for Cor-Incorporated #1
- guardrails: plugin MVP for policy enforcement #4
docs/ai-guardrails/migration/claude-code-skills-inventory.md
Activity
terisuke commentedon Apr 3, 2026
This issue is in the now-required set for the MVP floor tracked in #16. It should stay focused on safe local-runtime workflows and should not silently absorb later CI or broader migration scope.
terisuke commentedon Apr 3, 2026
Starting implementation from the source canon and the local brief. Scope for this PR: guarded primary agent, read-only review subagent, and explicit implement/review/ship/handoff commands only. CI authority and broader migration stay out of scope per #16.
feat: add safe agents and commands (#5)
feat: add safe agents and commands (#5) (#18)