feat(orb): detect APR repo-transfer acceptance/expiry via scheduled poll

[rsnetworkinginc]

Summary

Closes #7741 — initiateAprRepoTransfer returns { initiated: true } on a 2xx, but GitHub repo transfers are asynchronous and acceptance-gated; nothing detected whether a pending transfer was accepted, declined, or expired. This implements the ratified poll-based detection (poll, not webhook — a transfer to a customer's own account moves the repo outside our App's installation, so no webhook is reliably deliverable).

What changed (src/orb/apr-repo-transfer.ts)

Detection/expiry are pure, injectable functions (the live loop is wired but v8 ignore'd, matching the existing relay-drain pattern), so the logic is unit-tested directly:

• probeAprRepoTransfer — GET /repos/{owner}/{repo} with the App installation token. 404 → accepted-and-departed (ownership moved, App lost access); 2xx whose owner.login is now the target → resolved-under-target (accepted); anything else → still pending.
• classifyAprRepoTransferOutcome — pure: resolved → accepted, departed → accepted_departed, else expired once now - initiatedAt >= 7 days (APR_REPO_TRANSFER_EXPIRY_MS), else pending.
• pollPendingAprRepoTransfers — per pending transfer: probe → classify → reconcile.

Pause reuse (deliverable 2): setAprRepoDispatchPaused toggles the existing per-repo agentPaused kill-switch via upsertRepositorySettings — no new mechanism. requestAprRepoTransfer freezes dispatch on initiation; the poll resumes on accepted-and-still-installed or expired/declined, and leaves accepted_departed alone (nothing left to dispatch).

Scheduling (reused, not reinvented): a new poll-apr-repo-transfers job type, hourly enqueue in src/index.ts gated behind a default-off LOOPOVER_APR_TRANSFER_POLL flag (keeps existing cron tests byte-identical), and a dispatch arm in src/queue/job-dispatch.ts. The pending-transfer store fail-empties (no-op) until #7664 persists rows — mirroring the merged #8001 loadAprIdeaCompletion precedent.

Tests (deliverable 3)

Extended test/unit/orb-apr-repo-transfer.test.ts in its existing style: accepted, accepted-departed, expired-at-7-days, still-pending-within-window, plus probe/pause/flag cases; added a flag-on enqueue test to test/unit/index.test.ts.

Verified locally on top of current main: npm run typecheck clean, @loopover/engine build clean, git diff --check clean, 82 tests pass, and 100% patch coverage — src/orb/apr-repo-transfer.ts and src/index.ts both report zero uncovered statements and zero uncovered branches. The job-dispatch.ts live-loop arm is v8 ignore'd (the injectable logic it binds is covered directly in the unit test).

[superagent-security]

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

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-22 17:23:39 UTC}

7 files · 1 AI reviewer · no blockers · readiness 95/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This implements poll-based detection of APR repo-transfer acceptance/expiry (#7741), reusing the existing agentPaused kill-switch for dispatch pause/resume, with all IO injected so classifyAprRepoTransferOutcome/probeAprRepoTransfer/pollPendingAprRepoTransfers are unit-tested directly. The feature is correctly flag-gated (LOOPOVER_APR_TRANSFER_POLL, default off) and the live-loop wiring in job-dispatch.ts is a no-op today since loadPendingAprRepoTransfers fail-empties until #7664 lands, matching the documented convergence pattern. One real design question: the module's own header states a transfer to a customer's own account moves the repo outside the App's installation, yet probeAprRepoTransfer's `resolved_under_target` (accepted, still-installed) branch requires a 2xx response with the App's installation token — for the actual customer-transfer flow this feature targets, that branch may never fire in practice (the App would 404 instead), making 'accepted' effectively unreachable for the flagship use case and worth a direct answer from the author rather than silent generality.

Nits — 4 non-blocking

• src/orb/apr-repo-transfer.ts:203-221 — confirm the `resolved_under_target` (still-installed) outcome is actually reachable for a customer-account transfer, given the module's own doc comment says that transfer type removes the App's installation access entirely; if it's dead for this feature, say so in the comment or drop the distinction.
• src/orb/apr-repo-transfer.ts:271 pollPendingAprRepoTransfers re-asserts setDispatchPaused(true) every poll pass for a still-pending transfer even though it was already paused at initiation — harmless but worth a one-line note on why the idempotent re-assert exists (clock skew / a settings row that drifted back to unpaused externally).
• Add a short note near probeAprRepoTransfer clarifying whether fetch's default redirect-follow behavior is relied upon to reach the new path's 2xx body, since that's the crux of the `resolved_under_target` detection.
• Consider a named constant for the repeated `404` check across initiateAprRepoTransfer-adjacent probes for consistency, though this matches existing conventions in the file.

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	#7741
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: 16 registered-repo PR(s), 5 merged, 0 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	rsnetworkinginc; Gittensor profile; 16 PR(s), 0 issue(s).
Improvement	✅ Minor	risk: clean · value: minor

Linked issue satisfaction

Addressed
The PR implements all three deliverables: a scheduled poll (via existing cron/queue infra, hourly, gated behind a flag) that probes GitHub and classifies pending transfers as accepted/accepted_departed/pending/expired at the ratified 7-day window, reuse of the existing per-repo agentPaused kill-switch to pause on initiation and resume on resolution, and unit tests covering accepted, expired, and s

Review context

• Author: rsnetworkinginc
• 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: 16 PR(s), 0 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Start here: Triage stale or unlinked PRs.

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

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.30%. Comparing base (2656eaa) to head (2804f59).
⚠️ Report is 3 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8005      +/-   ##
==========================================
- Coverage   91.92%   89.30%   -2.63%     
==========================================
  Files         741       98     -643     
  Lines       75957    22151   -53806     
  Branches    23048     3863   -19185     
==========================================
- Hits        69827    19781   -50046     
+ Misses       5037     2188    -2849     
+ Partials     1093      182     -911     

Flag	Coverage Δ
shard-1	58.58% <97.14%> (+1.78%)	⬆️
shard-2	6.39% <2.85%> (-46.29%)	⬇️
shard-3	42.08% <8.57%> (-9.17%)	⬇️

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

Files with missing lines	Coverage Δ
src/index.ts	96.15% <100.00%> (+0.03%)	⬆️
src/orb/apr-repo-transfer.ts	100.00% <100.00%> (ø)
src/queue/job-dispatch.ts	99.29% <ø> (-0.71%)	⬇️
src/types.ts	100.00% <ø> (ø)

... and 643 files with indirect coverage changes