docs: propose verifier-bound skill scorecards

[solomonneas]

Summary

• Defines multi-dimension skill scorecards whose eligibility and values come only from verifier receipt v2 fields.
• Adds a sharp subject, patch, work-owner, verifier, check-role, and failure-taxonomy trust boundary.
• Specifies bounded exploration, dual effectiveness and utility promotion, immediate routing demotion, strict cold start, and audit-only backfill.
• Decomposes the implementation into four new S/M issues plus existing outcome ledger: dual-criterion promotion (effectiveness AND utility retention) #503.

Current failure mode

scored_artifact_count is only the number of distinct artifact ids with parsed outcome rows. Capture is opt-in, rank is read-only, and only explicit brigade outcome reconcile --apply advances promotion state. The proposal does not treat the reported 1,601 existing rows as trustworthy training data.

Verification

• brigade work verify run --target . --command "git diff --cached --check" --command "brigade handoff lint --target . .claude/memory-handoffs/2026-07-26-skill-scorecards-planning.md" --capture brigade-work
• Receipt: 20260726-223449-work-verify-3713cb
• MiseLedger evidence: 68adffeab4dd404c3b95b47d
• Both commands exited 0.

Refs #502 and #503.

[coderabbitai]

Important

Review skipped

Auto reviews are limited based on label configuration.

🏷️ Required labels (at least one) (1)

• coderabbit-review

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository: escoffier-labs/coderabbit/.coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 4221d161-dd60-4e8c-ad23-24a6befc73fc

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch t3code/plan-skill-scorecards

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[solomonneas]

Blocking readiness on one stale dependency reference:

• The opening schema-contract section says chore(receipts): schema audit and canonicalization for SIEM-readiness #562 is pending and points implementers to origin/pr-562. chore(receipts): schema audit and canonicalization for SIEM-readiness #562 merged as ad5541f532abfc53ce480b5bcf2e23b5cca3b4de, and that remote ref does not exist.
• Update the reference to docs/receipt-schemas.md on main, state that chore(receipts): schema audit and canonicalization for SIEM-readiness #562 is merged, and rebase so CI validates the proposal against the merged contract.

I found no other issue in the scorecard decomposition or its four filed implementation slices.