feat(slop): issue-side slop triage (#533)

[JSONbored]

What

Extends the anti-slop pillar from PRs to issues (closes #533) — a deterministic, advisory-only maintainer triage signal for clearly low-effort issues. Issues have no gate, so these never block; they flag issues for the maintainer.

How

• src/signals/slop.ts — buildIssueSlopAssessment with two high-precision, conservative signals:
  • empty_issue_body — empty/whitespace body.
  • unfilled_issue_template — a body that reduces to nothing after stripping template scaffolding (HTML comments, headings, empty bullets/checkboxes), i.e. an opened-but-unfilled template.
  • Mutually exclusive; any real prose → clean, so a terse-but-real issue ("Typo on line 12") stays clean. Conservative by design — the bar is "clearly low-effort."
• processors — the issue webhook path computes issue slop when the repo opted in (slopGateMode != off) and appends the findings to the (maintainer-facing) issue advisory.
• MCP gittensory_check_issue_slop — pure local-metadata (title + body, no repo/auth scope), mirroring gittensory_check_slop_risk; returns slopRisk / band / findings + the issue rubric. The on-demand surface for a triage agent.

Tests

The detector (empty / unfilled / clean / mutual-exclusion + the direct standalone guards), the MCP tool (low-effort vs genuine issue), and the webhook wiring (opted-in repo gets the finding; default-off repo does not). Typecheck + UI lint + OpenAPI clean; 97% coverage held.

Closes #533.

[ghost]

Note

Gittensory Gate skipped

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

Signal	Result	Evidence	Action
Gate result	⚠️ Skipped	#729 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.

[superagent-security]

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

[ghost]

gittensory · advisory review

Reviewed 6 changed file(s) — two independent AI reviewers.

Suggested action: ✅ Safe to merge — both reviewers found no blocking issues.

Reviewer A · gpt-oss-120b — recommends ✅ merge
The PR introduces issue‑side slop triage: new assessment functions, a public MCP tool, integration into issue advisory processing, and comprehensive tests. The implementation follows existing patterns, respects the public/private boundary, and compiles cleanly.

Suggestions

• Add a brief comment to the exported constants (ISSUE_SLOP_RUBRIC_MARKDOWN, ISSUE_SLOP_WEIGHTS) for future maintainers.
• Consider adding a test case where the issue body is undefined to confirm the assessment treats it as empty.
• Verify that the advisory findings added by buildIssueSlopAssessment are never exposed via any public API endpoint.

Worth double-checking

• Ensure the new issue‑slop findings remain internal‑only and are not inadvertently rendered in public comment surfaces.
• Confirm that the added MCP tool respects input size limits and does not introduce performance regressions.

Reviewer B · mistral-small-3.1-24b-instruct — recommends ✅ merge
This PR adds a new feature for issue-side slop triage, introducing signals to identify low-effort issues. The implementation is thorough, with comprehensive tests and clear documentation. The changes are well-integrated into the existing codebase.

Suggestions

• Consider adding a comment in the buildIssueSlopAssessment function explaining why the emptyBodyFinding check is done before the unfilledTemplateFinding check.
• Ensure that the ISSUE_SLOP_RUBRIC_MARKDOWN is kept up-to-date if any new signals are added in the future.
• Add a test case for the edge scenario where the issue body contains only whitespace characters but is not entirely empty (e.g., multiple spaces or tabs).

Worth double-checking

• Verify that the new signals do not introduce any false positives or negatives in real-world scenarios.
• Ensure that the checkIssueSlop method handles edge cases where the input is malformed or incomplete.
• Double-check that the new feature does not inadvertently expose any sensitive information in public comments.