mission.md

Mission - Verification Gates

Goal

Implement verify(task_id, artifacts) as a pure deterministic function over scope report, rule report, feedback log, and diff, emitting one verification_report.json per task close-out.

Inputs

• Stub loaders for scope_report.json, rule_report.json, feedback_record.jsonl, and the diff
• The check table: acceptance ran, acceptance exited zero, scope clean, no null exits, all block-severity rules pass

Deliverables

• A pure verify(task_id, artifacts) -> VerdictReport
• A printer that shows per-check results and the final pass/fail
• Three demo scenarios written to disk: clean pass, scope creep, missing acceptance

Acceptance

• python3 code/main.py exits zero
• The clean-pass scenario reports passed: true; the other two report passed: false
• Each scenario writes a separate verification_report.json under outputs/verification/

Out of scope

• LLM-as-judge logic. The gate stays deterministic; qualitative judgment belongs to the reviewer in lesson 39.
• Signed override audit logs. The exercise prompts extend the gate that way.

References

• docs/en.md - full lesson
• code/main.py - reference implementation
• outputs/skill-verification-gate.md - extracted skill