[setchy]

Mirrors the checks indicator GitHub shows in the pull request list (e.g. ✓ 50/50).

• Fetches statusCheckRollup on the PR's head commit, using the per-state counts (checkRunCountsByState / statusContextCountsByState), so no pagination of individual checks is needed.
• New ChecksPill shows passed/total:
  • green ✓ when everything passed
  • red ✗ when any check failed
  • yellow • while checks are still running
• Tooltip and screen reader label: e.g. Checks: 48 passed, 1 failed, 1 pending.
• Placed just before the reviews pill.

State mapping

Outcome	Check run states	Status context states
Passed	SUCCESS, NEUTRAL, SKIPPED, COMPLETED	SUCCESS
Failed	FAILURE, CANCELLED, TIMED_OUT, ACTION_REQUIRED, STARTUP_FAILURE, STALE	FAILURE, ERROR
Pending	QUEUED, IN_PROGRESS, PENDING, WAITING	PENDING, EXPECTED

Regenerated GitHub GraphQL types via pnpm codegen.

[setchy]

Future idea: progress wheel for checks

Keeping the ✓ / ✗ / • icons with passed/total for now. An alternative we discussed for later, which would save space and match the sub-issue / task progress pills:

Multi-coloured segmented wheel, no number

• Reuse the SVG wheel from SubIssueProgressWheel, drawing one arc per outcome: green (passed), red (failed), yellow (pending), over the existing faint track.
• Drop the 50/50 text; exact counts stay in the tooltip and aria-label (e.g. Checks: 48 passed, 1 failed, 1 pending).

Things to solve

• At 12px the ring is ~38px around, so 1 failure out of 50 is under 1px of red. Give any non-zero failed/pending count a minimum segment length (~10% of the ring).
• Start the red segment at 12 o'clock so failures are seen first.
• Even then, a single failure is a subtler signal than ✗, so compare both on real notifications before switching.

Simpler variants considered:

• Single-colour wheel (colour by overall state), with or without the number.
• Current icons, but show only ✓ (no count) when everything passed.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
92.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud