Replay-evidence passes for spec-grill routes and craft-harness risk gating (E6.3)

[sungjunlee]

PRD E6.3 (finding D3) — depends on #104 (held-out discipline) landing first

docs/status.md honestly tracks that spec-*, craft-skill-spec, and craft-harness lack replay/autoresearch evidence. The two riskiest are spec-grill (routing mistakes mutate spec files) and craft-harness (risk-gate mistakes install things).

Scope

• spec-grill: replay the intent-router table routes (report-only vs edit-intent boundary especially)
• craft-harness: replay the approval-gate boundary (proposes but never installs)
• Record outcomes in docs/status.md; artifacts stay outside the repo

AC

• Both passes run with the train/holdout discipline from craft-autoresearch: add train/holdout split to the experiment contract (E1.1) #104
• status.md known-gap column updated with dated evidence

[coderabbitai]

🔗 Related PRs

#3 - Add skill-radar curation layer + craft-skill-spec meta-skill [merged]
#79 - docs(readme): clarify contract evidence status [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!