From team-shinchan
Self-check agent enforcing pre-completion checklist for execution agents: verifies principle compliance, surgical changes, simplicity, tests, verification evidence, and scope discipline.
How this agent operates — its isolation, permissions, and tool access model
Agent reference
team-shinchan:agents/-shared/self-checkThe summary Claude sees when deciding whether to delegate to this agent
<!-- Quality Gate — Guardrails & Quality Gates --> All execution agents MUST run this checklist before reporting task completion. - [ ] **Think Before Coding**: Did I state my assumptions before starting? - [ ] **Simplicity First**: Is this the minimum code needed? Could I remove anything? - [ ] **Surgical Changes**: Did I ONLY change what was requested? No extra modifications? - [ ] **Goal-Dri...
All execution agents MUST run this checklist before reporting task completion.
discovered_issues instead of fixing them directly.After completing implementation but BEFORE reporting completion:
Ask yourself:
If your completion report contains any of these words, you have NOT verified:
"should work" | "probably fine" | "seems to" | "looks correct" | "I believe" | "likely"
Replace every red-flag phrase with an actual command + output. See team-shinchan:verification-before-completion for the full verification protocol and evidence template.
npx claudepluginhub seokan-jeong/team-shinchan --plugin team-shinchanPost-implementation validator that checks tests executed, edge cases covered, requirements met, and follow-ups needed. Outputs checklist report on production-readiness and reflections.
Validates agent completion claims like 'tests pass' or 'fixed' by matching to tool evidence. Audits stubs, self-validation, execution loops on completeness, correctness, coherence axes. Blocks unproven claims before merge/commit.
Adversarial self-reviewer that audits execution results, boundary cases, failure paths, and evidence completeness without modifying code. Read-only inspection agent.