feat(signals): slop signal — duplicate-cluster membership (#563)

[oktofeesh1]

Closes #563 (epic #530). Resubmit of #944 with the review concern structurally removed.

What

Adds the duplicate-cluster-membership slop signal: fires when a PR sits in a high-risk collision cluster holding 2+ open pull requests (genuine overlapping/duplicate work).

Change since #944 — addresses the review note at the root

The prior PR was closed on Reviewer A's concern that slop.ts's CollisionReport type import "might break compilation" and that a test import was "unused" (both were false — CI was green). To remove the trigger entirely rather than argue it:

• The cluster determination now lives in engine.ts as a pure exported helper isPullRequestInDuplicateCluster(collisions, pullNumber), beside the collision types it reads.
• slop.ts no longer imports CollisionReport — it takes a precomputed inDuplicateCluster: boolean.
• The helper is unit-tested in signals.test.ts with a typed CollisionReport fixture (compile-checked, so a shape mismatch is impossible).

How (precise, zero added gate cost)

• The gate reuses the collision report it already builds for queue health — no extra DB load or O(n²) compute.
• The 2+-pull-request bar excludes a healthy issue↔its-own-PR pair (also high-risk), keeping this blocking signal false-positive-averse (feat(api): maintainer quality-dashboard service + endpoint #557 lesson).
• Weighted 15; static public-safe text; inert on the local lint surfaces.

Tests

Helper branches (high+2PR+member → true; 1-PR / non-high / non-member / empty → false) in signals.test.ts; the slop finding (flagged/not-flagged + combined-weight band → elevated) in slop.test.ts. Full npm run test:coverage green (branches 97.03%); new lines fully covered.

🤖 Generated with Claude Code

[ghost]

reviewbot — ⏳ Waiting for CI to finish before the gate review.

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[ghost]

Note

Gittensory Gate skipped

PR closed before full evaluation. No late first comment was created.

Signal	Result	Evidence	Action
Gate result	⚠️ Skipped	#945 is no longer open.	No action.

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by Gittensory, a quiet PR intelligence layer for OSS maintainers.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.62%. Comparing base (9cbdc6a) to head (7b7dc8e).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #945   +/-   ##
=======================================
  Coverage   96.62%   96.62%           
=======================================
  Files         110      110           
  Lines       14997    15006    +9     
  Branches     5419     5423    +4     
=======================================
+ Hits        14491    14500    +9     
  Misses        104      104           
  Partials      402      402           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.