fix(github): guard the NaN expiry in mintInstallationToken

[shin-core]

What & why

mintInstallationToken's local App-JWT path parsed GitHub's expires_at straight into the cache entry with no finiteness check:

const expiresAtMs = payload.expires_at ? Date.parse(payload.expires_at) : Date.now() + 50 * 60_000;

Date.parse returns NaN for a present-but-unparseable string, and a NaN expiresAtMs makes the cache-freshness comparison expiresAtMs - TOKEN_SAFETY_MARGIN_MS > Date.now() forever false. So a single malformed expires_at silently disables the installation-token cache: every job re-mints, re-triggering exactly the thundering-herd (and the GitHub secondary-rate-limit on the token endpoint) the cache exists to prevent.

The fix

Treat an unparseable expires_at identically to an absent one — fall back to Date.now() + 50 * 60_000:

const parsedExpiry = payload.expires_at ? Date.parse(payload.expires_at) : Number.NaN;
const expiresAtMs = Number.isFinite(parsedExpiry) ? parsedExpiry : Date.now() + 50 * 60_000;

• Well-formed expires_at → still exactly Date.parse(payload.expires_at) (byte-identical).
• Absent expires_at → still the Date.now() + 50 * 60_000 fallback.
• Unparseable expires_at → now the same finite fallback instead of NaN.

No cache-lifetime constant changes (TOKEN_SAFETY_MARGIN_MS, the 50-minute fallback, all unchanged).

Tests (test/unit/github-app.test.ts)

• Regression (named for the bug): after a mint whose expires_at is unparseable, a second createInstallationToken for the same installation issues no further /access_tokens request — asserted by the mint-call count (mints === 1), not just the token string (fails on main, which re-mints).
• Well-formed expires_at: the cached token is reused on the second call (mint count 1) — unchanged behaviour.

Validation

• npm run typecheck clean for this file (the only local errors are a pre-existing missing-release-please-dep phantom in unrelated release-tooling, present on bare main); the github-app suite green.
• Diff coverage on src/github/app.ts is 100% (both Number.isFinite arms and both expires_at? arms).
• git diff --check <base> HEAD clean; single-file source change + its test.

Closes #10026

[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: 2026-07-31 06:36:34 UTC}

2 files · 1 AI reviewer · no blockers · readiness 98/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This is a correct, narrowly-scoped bug fix: the local App-JWT path in mintInstallationToken (src/github/app.ts:361-369) parsed a present-but-unparseable expires_at into NaN via Date.parse, and NaN silently defeats the cache-freshness check (expiresAtMs - MARGIN > Date.now() is forever false), forcing a re-mint on every call. The fix correctly falls through to Number.isFinite(parsedExpiry) ? parsedExpiry : Date.now() + 50*60_000, matching the existing absent-expires_at fallback exactly and leaving the well-formed path byte-identical. The included regression test directly exercises the real bug by stubbing expires_at as 'not-a-date' and asserting mint count stays at 1, which is a real (not fabricated) test of the fixed path, and a second test confirms well-formed behavior is unchanged.

Nits — 4 non-blocking

• The `50 * 60_000` fallback duplicates the literal already used on the absent-expires_at branch just above it in the pre-existing code — consider extracting a named constant (e.g. `DEFAULT_TOKEN_TTL_MS`) shared by both branches, per the external brief's magic-number flag.
• PR description doesn't explicitly link/close an eligible open issue in the repo-required format, though the code comment and test both reference github(app): guard the NaN expiry in mintInstallationToken so an unparseable expires_at cannot disable the installation-token cache #10026 — worth confirming that issue is open and this PR is tied to it per the repo's issue-linking requirement.
• src/github/app.ts:368 — factor `50 * 60_000` into a shared constant used by both the absent and unparseable-expiry fallback paths to avoid future drift between them.
• Consider a follow-up log line (warn-level) when the unparseable-expiry fallback path is hit, since a malformed expires_at from GitHub is itself signal worth surfacing, distinct from the silent NaN behavior being fixed here.

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	#10026
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: 64 registered-repo PR(s), 50 merged, 0 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	shin-core; Gittensor profile; 64 PR(s), 0 issue(s).
Improvement	✅ Minor	risk: clean · value: minor · LLM: moderate

Linked issue satisfaction

Addressed
The diff mirrors the broker-client guard exactly (Number.isFinite check with the same 50*60_000 fallback), preserving byte-identical behavior for well-formed and absent expires_at, and adds both a regression test asserting mint-call count for the unparseable case and a test confirming the well-formed path is unchanged.

Review context

• Author: shin-core
• 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: 64 PR(s), 0 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.

🧪 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 79.66%. Comparing base (eca3b61) to head (b82afd9).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #10074      +/-   ##
==========================================
+ Coverage   79.57%   79.66%   +0.09%     
==========================================
  Files         282      283       +1     
  Lines       58664    58964     +300     
  Branches     6842     6955     +113     
==========================================
+ Hits        46682    46976     +294     
- Misses      11694    11695       +1     
- Partials      288      293       +5     

Flag	Coverage Δ
backend	98.00% <100.00%> (?)

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

Files with missing lines	Coverage Δ
src/github/app.ts	98.00% <100.00%> (ø)

[loopover-orb]

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