feat(miner-governor): self-plagiarism throttle across the miner's own repos

[andriypolanski]

Closes #2345

Summary

Adds a pure selfPlagiarismCheck() consulted before open_pr, comparing a prospective PR's diff fingerprint against the miner's own recent submission history. Near-duplicates across repos are throttled using the same claim-time / earliest-wins election as isDuplicateClusterWinnerByClaim. Denials are recorded to the governor ledger with the flagged prior submission referenced.

Changes

Area	Change
packages/gittensory-engine/src/governor/self-plagiarism.ts	selfPlagiarismCheck, fingerprintSimilarity, ledger event builder, config resolver
packages/gittensory-engine/src/miner-goal-spec.ts	selfPlagiarism.similarityThreshold from .gittensory-miner.yml (default 0.85)
packages/gittensory-engine/src/index.ts	Export self-plagiarism module
packages/gittensory-miner/lib/governor-open-pr.js	evaluateOpenPrSelfPlagiarism — check + append-only ledger write
packages/gittensory-miner/docs/miner-goal-spec.md	Document selfPlagiarism field
packages/gittensory-miner/schema/miner-goal-spec.schema.json	JSON Schema for selfPlagiarism
.gittensory-miner.yml.example	Document selfPlagiarism block
Tests	self-plagiarism.test.ts, miner-governor-open-pr.test.ts, miner-goal-spec parser/doc updates

Behavior

• Distinct PR → allowed
• Near-duplicate across repos (similarity ≥ threshold) when a prior claimed first → throttled + ledger row references matched prior
• Missing candidate fingerprint or submittedAt → denied (fail closed)
• Near-duplicate prior missing submittedAt → denied (ambiguous election)
• Earliest near-duplicate claimant → allowed (mirrors duplicate-cluster winner election)

Test plan

• selfPlagiarismCheck — distinct PR passes
• selfPlagiarismCheck — near-duplicate across repos throttled
• selfPlagiarismCheck — missing/ambiguous fingerprint data fails closed
• selfPlagiarismCheck — earliest claimant wins among near-duplicates
• buildSelfPlagiarismGovernorLedgerEvent — references matched prior in payload
• evaluateOpenPrSelfPlagiarism — persists throttled denial to governor ledger
• parseMinerGoalSpec — parses and normalizes selfPlagiarism.similarityThreshold
• Miner goal spec docs/schema/example include selfPlagiarism
• npm run test:unit -- test/unit/self-plagiarism.test.ts test/unit/miner-governor-open-pr.test.ts test/unit/miner-goal-spec-doc.test.ts test/unit/miner-goal-spec-parser.test.ts
• npm run test:ci (full gate before push)

Notes

• Composes with the Governor open_pr chokepoint (maintainer: wire the fail-closed Governor chokepoint before every write action #2340) — this issue delivers the pure check + ledger recording helper, not the full chokepoint.
• Threshold is configurable via .gittensory-miner.yml → selfPlagiarism.similarityThreshold; built-in default is 0.85.
• Rationale vocabulary stays consistent with duplicate-cluster signals (near_duplicate_self_plagiarism).

Test plan

• Unit tests for check, ledger event builder, governor open_pr integration, and goal-spec parsing
• npm run test:ci

[superagent-security]

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

[loopover-orb]

Tip

🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩

✅ Gittensory review result - approve/merge recommended

_{Review updated: 2026-07-11 06:32:41 UTC}

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

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This adds a pure, deterministic self-plagiarism throttle (`selfPlagiarismCheck`) that Jaccard-compares a prospective PR's diff fingerprint against the miner's own recent submission history, reusing the existing claim-time election from `isDuplicateClusterWinnerByClaim` so exactly one near-duplicate claimant survives; it wires config through `.gittensory-miner.yml`'s new `selfPlagiarism.similarityThreshold` (default 0.85) and records verdicts to the governor ledger via `evaluateOpenPrSelfPlagiarism`. The fail-closed handling of missing/ambiguous fingerprint and timing data is correct and thoroughly exercised, including several deliberately-forced edge cases (threshold-0 zero-similarity tie, `resolveDuplicateClusterWinnerNumber` returning null) that are genuinely reachable, not fabricated. The PR closes #2345, stays scoped to the stated feature, and ships matching schema/docs/example updates alongside real integration tests that exercise the actual sqlite-backed governor ledger rather than mocks.

Nits — 5 non-blocking

• `selfPlagiarismCheck` doesn't restrict comparison to a *different* repo despite the docs/PR title framing this as throttling duplication 'across repos' — a same-repo resubmission after a closed/rejected PR (e.g. addressing review feedback) would also be flagged as a near-duplicate unless the caller pre-filters `recentOwnSubmissions`; worth confirming the intended caller-side filtering for `governor-open-pr.js:evaluateOpenPrSelfPlagiarism`.
• codecov/patch reports below the 99% target (73.80% and 98.82% across the listed runs) — worth checking which changed branches in `self-plagiarism.ts` or `miner-goal-spec.ts` are under-covered before merge.
• `validate-code` and `validate` are in the FAILED CI list with no visible cause in this diff; confirm these aren't type/lint regressions from the new `governor/self-plagiarism.ts` module before merge.
• In `buildSelfPlagiarismGovernorLedgerEvent` (self-plagiarism.ts), the `matched` prior used in the ledger payload is `bestMatch` (highest-similarity prior) rather than the actual election winner from `resolveDuplicateClusterWinnerNumber` when they diverge — the fallback chain finds a submission matching `winner`'s number only if `bestMatch` is undefined, so the referenced 'flagged prior' in a ledger row could differ from the true cluster winner in a three-way tie; consider a comment clarifying this is intentional (most-similar, not necessarily earliest).
• Add a short comment on `selfPlagiarismCheck` (self-plagiarism.ts:126) clarifying whether `recentOwnSubmissions` is expected to be pre-filtered to cross-repo entries only, since the current implementation is repo-agnostic.

Signal	Result	Evidence
Code review	✅ No blockers	1 reviewer
Linked issue	✅ Linked	#2345
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	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 140 registered-repo PR(s), 95 merged, 25 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	andriypolanski; Gittensor profile; 140 PR(s), 25 issue(s).
Gate result	✅ Passing	No configured blocker found.
Improvement	✅ Minor	risk: clean · value: minor — Code changes are accompanied by test evidence. LLM value judgment: moderate — The diff adds a scoped, well-tested governor safeguard (self-plagiarism throttle) tied to a specific linked issue, reusing existing duplicate-winner election logic rather than reinventing it, which is a meaningful but narrow addition to the governor's open_pr chokepoint.

Review context

• Author: andriypolanski
• 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: 140 PR(s), 25 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Keep the PR focused and include validation evidence before maintainer review.

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

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.27%. Comparing base (b94589c) to head (2ebcc8f).
⚠️ Report is 5 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4972      +/-   ##
==========================================
+ Coverage   94.26%   94.27%   +0.01%     
==========================================
  Files         448      449       +1     
  Lines       38900    38985      +85     
  Branches    14172    14211      +39     
==========================================
+ Hits        36669    36754      +85     
  Misses       1574     1574              
  Partials      657      657              

Flag	Coverage Δ
shard-1	46.94% <87.05%> (+0.11%)	⬆️
shard-2	33.21% <2.35%> (-0.11%)	⬇️
shard-3	31.48% <10.58%> (+0.06%)	⬆️
shard-4	33.08% <2.35%> (-0.07%)	⬇️
shard-5	33.56% <42.35%> (+0.02%)	⬆️
shard-6	45.19% <20.00%> (-0.06%)	⬇️

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

Files with missing lines	Coverage Δ
.../gittensory-engine/src/governor/self-plagiarism.ts	100.00% <100.00%> (ø)
packages/gittensory-engine/src/miner-goal-spec.ts	97.91% <100.00%> (+0.26%)	⬆️

🚀 New features to boost your workflow:

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

[loopover-orb]

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