feat(mcp): REST route + CLI mirror for loopover_get_eligibility_plan

[dhgoal]

Summary

Closes #6621

loopover_get_eligibility_plan existed only on the remote MCP server (src/mcp/server.ts) — the one tool in the preview / explain_breakdown / eligibility_plan trio with no REST route and no local CLI mirror. This adds both, modelled exactly on loopover_explain_score_breakdown's existing pair.

• POST /v1/scoring/eligibility-plan (src/api/routes.ts), placed immediately after explain-breakdown and structured identically: same scorePreviewSchema, same repo/snapshot/evidence/contributorIssues fetch, same buildScorePreview(...). It returns deriveEligibilityPlan(preview) (reused as-is from src/services/eligibility-plan.ts, not reimplemented) and — like /v1/scoring/preview, and matching loopover_get_eligibility_plan's own MCP handler — treats contributorLogin as optional (conditional gate), not the unconditional requirement explain-breakdown uses.
• loopover_get_eligibility_plan stdio tool in packages/loopover-mcp/bin/loopover-mcp.js, plus a STDIO_TOOL_DESCRIPTORS entry (category discovery, matching the server's MCP_TOOL_CATEGORIES). The local branch-metadata→request-body assembly it shares with loopover_explain_score_breakdown is factored into a buildLocalScoreRequestBody helper both call, per the issue's preferred option, so the two can't drift — the only difference is the apiPost path.

Scope

• The PR title follows type(scope): short summary Conventional Commit format.
• This PR is focused and does not mix unrelated backend, UI, MCP, docs, dependency, and deploy changes.
• This follows CONTRIBUTING.md and does not reintroduce GitHub Pages, VitePress, site/, or CNAME.
• I linked a currently open issue this PR resolves (Closes #6621).

Validation

• git diff --check
• npm run actionlint
• npm run typecheck (see note — type-safe by construction; local root typecheck OOMs)
• npm run test:coverage (the changed route's every branch is covered — see below)
• npm run test:workers
• npm run build:mcp
• npm run test:mcp-pack
• npm run ui:openapi:check
• npm run ui:lint
• npm run ui:typecheck
• npm run ui:build
• npm audit --audit-level=moderate
• New or changed behavior has unit/integration tests for new branches, fallback paths, and sanitizer boundaries

If any required check was skipped, explain why:

• Route coverage (the codecov/patch-gated surface). Ran the affected integration tests green: test/integration/api.test.ts "serves deterministic signal endpoints" now exercises the new route's authorized path (contributorLogin present → 200 + {eligible, branchEligibilityStatus, blockers, cleanupPaths}), the optional-login path (no contributorLogin → 200, the branch that differs from explain-breakdown), and the invalid-body 400 (added to the parametrized invalid_scoring_preview_request list). The contributor-gate 403 branch is covered in test/integration/routes-errors.test.ts, mirroring the existing victimScorePreview case (a session whose actor ≠ the requested contributor). That is every branch of the new route.
• CLI coverage. packages/loopover-mcp/bin/loopover-mcp.js is not in vitest.config.ts's coverage.include, so it isn't codecov-gated. Its registration, description, and category are validated by the existing test/unit/mcp-cli-tools.test.ts (which boots the real stdio server and asserts descriptors match registered tools) — passing with the new tool. The request-body assembly is now the shared buildLocalScoreRequestBody helper, exercised through loopover_explain_score_breakdown's unchanged path; the sole delta is the apiPost endpoint string.
• Typecheck. The route is byte-identical to the adjacent, type-checked explain-breakdown route except the terminal call; explainScoreBreakdown and deriveEligibilityPlan both take the same ScorePreviewResult, so passing the shared preview is type-safe by construction. The repo's root typecheck OOMs locally; an isolated tsc surfaces no error at the changed lines (only pre-existing ambient-type/config artifacts common to the whole file). CI runs the full typecheck.
• OpenAPI. src/openapi/spec.ts is hand-curated and does not list explain-breakdown (or preview's siblings); the new route follows that precedent, so ui:openapi:check needs no regen.
• Prettier is not run by CI (no format/format:check script, absent from test:ci); all four touched files are already non-prettier-clean on main, so no formatting churn was introduced.

Safety

• No secrets, wallet details, hotkeys, coldkeys, user PATs, private keys, raw trust scores, private rankings, or private maintainer evidence are exposed.
• Public GitHub text stays sanitized and low-noise.
• Auth, cookie, CORS, GitHub App, Cloudflare, or session changes include negative-path tests.
• API/OpenAPI/MCP behavior is updated and tested where needed.
• UI changes use live API data or real empty/error/loading states, not production mock/demo fallbacks.
• Visible UI changes include a UI Evidence section.
• Public docs/changelogs are updated where needed.

The new route inherits the exact contributor gate of its siblings: requireContributorAccess is applied when a contributorLogin is supplied (403 forbidden_contributor on a mismatch — tested), and deriveEligibilityPlan returns only public-safe blocker/cleanup language (no scores or private counts) by construction.

Notes

Supersedes #6660 (auto-closed): that revision added the stdio tool but missed the hardcoded tool-count invariant in test/unit/mcp-tool-rename-aliases.test.ts (60→61). That one test failing in validate-tests (3) also aborted that shard's Codecov upload, which is why the same root cause surfaced as both a test failure and a codecov/patch drop. Fixed here; the count assertions (and their it titles + running comment) now read 61, verified locally (10/10).

Notes

deriveEligibilityPlan and the remote loopover_get_eligibility_plan tool are unchanged; this only adds the missing HTTP + CLI reach. The shared buildLocalScoreRequestBody extraction removes the copy-paste the issue flagged, so explain_score_breakdown and get_eligibility_plan stay in lockstep.

Closes #6621

[superagent-security]

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

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.65%. Comparing base (58ab17e) to head (0c87556).
⚠️ Report is 8 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #6692   +/-   ##
=======================================
  Coverage   93.65%   93.65%           
=======================================
  Files         682      682           
  Lines       68112    68126   +14     
  Branches    18708    18713    +5     
=======================================
+ Hits        63788    63802   +14     
  Misses       3347     3347           
  Partials      977      977           

Flag	Coverage Δ
shard-1	43.94% <7.14%> (+0.04%)	⬆️
shard-2	36.94% <7.14%> (-0.03%)	⬇️
shard-3	32.47% <28.57%> (+0.01%)	⬆️
shard-4	34.66% <7.14%> (+0.42%)	⬆️
shard-5	31.67% <7.14%> (+0.29%)	⬆️
shard-6	45.39% <92.85%> (-0.36%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/api/routes.ts	94.60% <100.00%> (+0.03%)	⬆️

... and 1 file with indirect coverage changes

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-16 19:46:11 UTC}

5 files · 1 AI reviewer · no blockers · readiness 77/100 · CI green · unknown

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This PR adds a REST route (POST /v1/scoring/eligibility-plan) and a stdio CLI/MCP tool (loopover_get_eligibility_plan) mirroring the existing explain-breakdown pattern, and factors the shared local-branch-metadata assembly into a new buildLocalScoreRequestBody helper used by both tools. The route reuses deriveEligibilityPlan and buildScorePreview exactly as the existing preview/explain-breakdown routes do, correctly makes the contributor gate conditional (matching /v1/scoring/preview), and is backed by integration tests covering the 200 path, the optional-contributor path, the 403 forbidden-contributor gate, and the invalid-request 400 path. The tool-count bump test (61→62) is updated consistently with the new registration.

Nits — 4 non-blocking

• src/api/routes.ts:2121 — the extracted `400` status/`invalid_scoring_preview_request` error shape duplicates the pattern used by every other scoring route; consider a shared response helper if this recurs again, though it's consistent with existing style.
• The size-smell flags on loopover-mcp.js and routes.ts reflect pre-existing file size, not growth introduced by this diff specifically — worth tracking separately but not a blocker here.
• packages/loopover-mcp/bin/loopover-mcp.js — `buildLocalScoreRequestBody`'s destructured `estimatedSourceLines` fallback math (`diff.changedLineCount - diff.testFiles.length`) is copied verbatim from the pre-existing explain-breakdown logic; fine since it's unchanged behavior, but worth a shared comment if a third caller appears.
• Consider adding a unit test directly for `buildLocalScoreRequestBody` in isolation (not just via the two tools that call it) to pin its contract now that it's shared.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#6621
Related work	⚠️ 1 scoped overlap	Top overlaps are listed below; lower-confidence bulk is hidden.
Change scope	❌ 8/20	High review scope from cached public metadata (1 linked issue).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 128 registered-repo PR(s), 75 merged, 36 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	dhgoal; Gittensor profile; 128 PR(s), 36 issue(s).
Improvement	✅ Minor	risk: clean · value: minor

Review context

• Author: dhgoal
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: not available
• Official Gittensor activity: 128 PR(s), 36 issue(s).
• Related work: Titles/paths share 7 meaningful terms. (issue #6593, issue #6621)

Contributor next steps

• Start here: Review top overlaps.
• Then work through the remaining 3 steps in the Signals table above.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

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

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

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.