Release Evidence Pack

Use .claude/rules/evidence-schema.yml as the output schema and .claude/rules/synergy-contract.yml for specialist routing.

Workflow

1. Record changed files and classify runtime, test, docs, memory, and .claude impacts.
2. Run or collect evidence for python -m compileall -q src/, targeted pytest, full pytest when feasible, and .claude config validation when relevant. When .claude/**, CLAUDE.md, or AGENTS.md changed, include the live-smoke checklist from .claude/skills/claude-config-audit/references/live-smoke-template.md.
3. Summarize security checks, including secrets scan status and dependency audit status if dependencies changed.
4. List at least three edge cases considered.
5. State whether memory/ files changed and why.
6. Ask routed specialists to emit a fenced yaml block matching .claude/rules/evidence-schema.yml; when shell execution is available, validate captured blocks with python .claude/hooks/validate-evidence-payload.py <payload>.
7. Produce a GO/NO-GO recommendation with blockers separated from recommendations.

Output Sections

• Scope
• Commands and exact outcomes
• Specialist findings
• Edge cases
• Memory updates
• Remaining risk
• Final verdict

The final pack must include verdict, scope_reviewed, source_of_truth, commands_run, findings, edge_cases, handoffs, and memory_update fields from .claude/rules/evidence-schema.yml, plus validate-evidence-payload.py PASS/WARNING/FAIL status for specialist evidence when collection was possible.