feat(engine): wire historical replay into Phase 7 calibration loop

[andriypolanski]

Summary

Closes #3014.

Adds the pure engine contract for Phase 7 calibration loop wiring in @jsonbored/gittensory-engine. The historical-replay composite score (from objective-anchor, pairwise-judge, and structured gate-verdict inputs) is combined alongside the live pr_outcome-derived accuracy signal instead of silently replacing it. The loop tracks a combined calibration-accuracy metric anchored to the documented 62% baseline, records provenance per source, recommends replay-run cadence, and fail-closes autonomy-level increases when the replay harness is missing, stale, degraded, unavailable, or below the configured threshold.

Changed files

File	Change
packages/gittensory-engine/src/phase7-calibration-loop.ts	New pure module: config resolution, pr_outcome accuracy derivation, replay freshness/cadence, combined metric, autonomy gating, audit markdown.
packages/gittensory-engine/test/phase7-calibration-loop.test.ts	Unit/regression tests for config, combine, fail-closed holds, baseline anchoring, and markdown rendering.
packages/gittensory-engine/src/index.ts	Export Phase 7 calibration loop APIs.
packages/gittensory-engine/README.md	Document config surface and usage example.

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 an issue, or this is small enough that the summary explains why an issue is not needed.

Linked issue: #3014.

Validation

• git diff --check
• npm run actionlint
• npm run typecheck
• npm run test:ci — run before push (~8 min).
• npm audit --audit-level=moderate

[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 20:40:12 UTC}

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

⏸️ Suggested Action - Manual Review

Review summary
The new Phase 7 engine module is coherent: it resolves the opt-in config, derives the live pr_outcome metric from raw counters, combines it with a fresh healthy replay score, and fail-closes autonomy increases when replay evidence is missing or rejected. The tests exercise the main enabled/disabled, stale/degraded, threshold, export, and markdown paths, and the barrel export is consistent with the new module. I do not see a reachable correctness defect in the provided changed files.

Nits — 5 non-blocking

• nit: packages/gittensory-engine/src/phase7-calibration-loop.ts:52 defines `PrOutcomeCalibrationInput.hold` but `computePrOutcomeCalibrationAccuracy()` never reads it, so either document that held outcomes are intentionally excluded from the decided accuracy denominator or remove the field from this pure contract.
• nit: packages/gittensory-engine/src/phase7-calibration-loop.ts:210 reports all malformed top-level alias values as `miner.calibration.*` warnings, which can send operators to the wrong config path when they used `calibration.*`; consider including the resolved source path in each warning.
• nit: packages/gittensory-engine/src/phase7-calibration-loop.ts:292 treats `harnessStatus: "missing"` the same as an otherwise healthy/fresh replay for scheduling if `lastReplayObservedAt` is present; if callers can know the harness is missing independently of the last run timestamp, make that an explicit due reason.
• packages/gittensory-engine/src/phase7-calibration-loop.ts:52: add a test for a nonzero `hold` counter and assert the intended behavior, because that is the only pr_outcome field whose contribution is not obvious from the helper name.
• packages/gittensory-engine/src/phase7-calibration-loop.ts:210: change warning construction to name `miner.calibration` or `calibration` based on which raw value actually won the overlay.

Signal	Result	Evidence
Code review	✅ No blockers	1 reviewer
Linked issue	✅ Linked	#3014
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: 110 registered-repo PR(s), 74 merged, 5 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	andriypolanski; Gittensor profile; 110 PR(s), 5 issue(s).
Gate result	✅ Passing	No configured blocker found.

Review context

• Author: andriypolanski
• 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: 110 PR(s), 5 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.