feat(miner): add pairwise calibration scoring

[joaovictor91123]

Summary

• Adds deterministic objective-anchor feature extraction/scoring for replayed plans versus revealed history, including per-item history aggregation and Markdown audit rendering.
• Adds the pairwise calibration scorer for order-swapped judge attempts, retry caps, instability metrics, and composite scoring with objective-anchor results.
• Fixes feat(miner): pairwise-judge scoring layer with order-swap bias defense #3013

Scope

• The PR title follows type(scope): short summary Conventional Commit format, for example fix(api): restore profile access checks.
• 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 an issue, or this is small enough that the summary explains why an issue is not needed.

Validation

• git diff --check HEAD~1..HEAD
• npm run actionlint
• npm run typecheck
• npm run test:coverage locally; codecov/patch requires >=99% coverage of the lines AND branches you changed (aim for 100% on your diff so CI variance does not fail near the threshold). Global coverage is a non-blocking trend with a loose 90% backstop, not the gate.
• 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:

• npm test --workspace @jsonbored/gittensory-engine passed after rebase: 102 tests.
• npm run build:miner passed after rebase.
• npm run test:coverage and npx vitest run --coverage --reporter=dot both timed out locally on Windows after long runs and ended with EPIPE after the session killed the output pipe, not with a reported assertion failure.
• npm run test:mcp-pack failed locally on Windows/Node 24 in scripts/check-mcp-package.mjs while writing an undefined spawn diagnostic for npm pack; this feature does not touch MCP packaging.
• UI checks were not run because this PR changes only the engine package and its package README.

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, low-noise, and does not imply compensation guarantees or optimization tactics.
• 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 below with JPG/JPEG or PNG screenshots arranged as organized, captioned, clickable thumbnails. SVG screenshots are not used as review evidence. Review-only screenshots or recordings are not committed to the repository.
• Public docs/changelogs are updated where needed; changelogs are only edited for release-prep PRs.

UI Evidence

N/A - no visible UI, frontend, or extension change.

Notes

• The engine package owns the deterministic scoring and stability logic. The miner runtime can supply order-swapped judge attempts from its own model invocation layer.

[superagent-security]

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

[loopover-orb]

Warning

🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨

⏸️ Gittensory review result - manual review recommended

_{Review updated: 2026-07-04 17:15:34 UTC}

6 files · 1 AI reviewer · no blockers · readiness 80/100 · CI green · clean

⏸️ Suggested Action - Manual Review

Review summary
The PR adds a pure objective-anchor scorer, Markdown audit rendering, and a pairwise calibration combiner with public exports and direct tests. The core scoring paths are deterministic, covered, and the pairwise order-swap handling is coherent for stable replay/revealed/tie outcomes. I do not see a reachable correctness break in the visible diff, but there are a couple of contract edges worth tightening before this becomes a calibration API callers depend on.

Nits — 5 non-blocking

• nit: packages/gittensory-engine/src/objective-anchor.ts:378 aggregates per-item `unknown` change kinds alongside known kinds, so one sparse history item can dilute an otherwise exact known-kind match; either drop `unknown` when any concrete kind is present in `combineFeatures`, or document that missing metadata intentionally penalizes aggregate history scores.
• nit: packages/gittensory-engine/src/pairwise-calibration.ts:91 accepts non-finite `maxAttempts` values, so `Infinity` yields `maxAttempts: Infinity` and an unstable finite attempt list reports `exhausted: false`; clamp this the same way weights are normalized or require a finite positive integer.
• nit: packages/gittensory-engine/src/pairwise-calibration.ts:126 uses `input.objectiveAnchor.score` as-is for `ObjectiveAnchorScore` inputs while numeric inputs are rounded/clamped; normalize both branches unless the API intentionally trusts score objects as already canonical.
• packages/gittensory-engine/src/objective-anchor.ts:191 should consider filtering `unknown` out of the aggregated `changeKinds` when concrete kinds exist, e.g. by applying that rule inside `combineFeatures`.
• packages/gittensory-engine/src/pairwise-calibration.ts:89 should add a small `finitePositiveInteger` helper for `maxAttempts` and cover `0`, negative, `NaN`, and `Infinity` in tests.

Signal	Result	Evidence
Code review	✅ No blockers	1 reviewer
Linked issue	✅ Linked	#3013
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (1 linked issue).
Validation posture	❌ 5/25	Preflight is holding this PR: the review lane is unavailable, so it is not ready for automated review.
Contributor workload	✅ 10/10	Author activity: 90 registered-repo PR(s), 49 merged, 8 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	joaovictor91123; Gittensor profile; 90 PR(s), 8 issue(s).
Gate result	✅ Passing	No configured blocker found.

Review context

• Author: joaovictor91123
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository registration is not available in the local Gittensory cache.
• Public profile languages: not available
• Official Gittensor activity: 90 PR(s), 8 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Await review-lane availability.
• Refresh registry data or choose a registered active repo.

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.

_{🟩 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 Gittensory, a quiet PR intelligence layer for OSS maintainers.

• Re-run Gittensory review

[loopover-orb]

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