Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Triage GitHub notifications and issue/PR queues.
.claude/skills/notque-github-notification-triage/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -42% | 0% |
Fetch, classify, and report on GitHub notifications. The script does the heavy lifting — this skill orchestrates invocation and presents results.
bash# Report-only (default): show what needs attention, no modifications python3 scripts/github-notification-triage.py # Mark informational notifications as read after reporting python3 scripts/github-notification-triage.py --mark-read # Save report to ~/.claude/reports/notifications/ python3 scripts/github-notification-triage.py --save # Cron/scheduled mode: auto-clear noise and save report python3 scripts/github-notification-triage.py --mark-read --save
Run report-only by default:
bashpython3 scripts/github-notification-triage.py
Display the script output directly to the user. The report classifies notifications into:
If the user responds with any of the following, re-run with --mark-read:
bashpython3 scripts/github-notification-triage.py --mark-read
Confirm how many notifications were marked read after the run completes.
When invoked on a schedule (no interactive user), use both flags to auto-clear and persist the report:
bashpython3 scripts/github-notification-triage.py --mark-read --save
| Signal | Load These Files | Why | |---|---|---| | triaging an issue or pull-request queue | item-card-contract.md | Defines the required per-item report card. |
When the task is triaging a queue of issues or PRs (not just notifications), report one card per item using the contract in references/item-card-contract.md:
git status --short empty) before any local work on an item.| Code | Meaning | |------|---------| | 0 | Success | | 1 | Error (auth failure, API error, script not found) |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 3,068 | 6,121 | +100% | 1 | 1 | 0% | 511 | 936 | +83% | 0 | 0 | — |
case-01 | fail→fail | 7,891 | 3,875 | -51% | 1 | 1 | 0% | 1,425 | 819 | -43% | 0 | 0 | — |
case-02 | fail→fail | 13,406 | 4,489 | -67% | 1 | 1 | 0% | 2,635 | 906 | -66% | 0 | 0 | — |
case-04 | fail→fail | 3,642 | 4,168 | +14% | 1 | 1 | 0% | 525 | 800 | +52% | 0 | 0 | — |
case-05 | fail→pass | 5,633 | 3,471 | -38% | 1 | 1 | 0% | 856 | 856 | 0% | 0 | 0 | — |
case-06 | fail→fail | 11,407 | 4,457 | -61% | 1 | 1 | 0% | 1,619 | 827 | -49% | 0 | 0 | — |
case-07 | fail→pass | 17,368 | 3,220 | -81% | 1 | 1 | 0% | 3,313 | 1,194 | -64% | 0 | 0 | — |
case-08 | fail→fail | 8,663 | 2,434 | -72% | 1 | 1 | 0% | 1,463 | 986 | -33% | 0 | 0 | — |
case-09 | fail→pass | 8,541 | 5,649 | -34% | 1 | 1 | 0% | 1,372 | 1,008 | -27% | 0 | 0 | — |
case-10 | fail→pass | 6,305 | 1,894 | -70% | 1 | 1 | 0% | 950 | 906 | -5% | 0 | 0 | — |
case-11 | fail→fail | 6,490 | 1,899 | -71% | 1 | 1 | 0% | 932 | 922 | -1% | 0 | 0 | — |
case-12 | fail→pass | 11,100 | 2,175 | -80% | 1 | 1 | 0% | 1,769 | 1,031 | -42% | 0 | 0 | — |
case-13 | fail→pass | 6,884 | 1,993 | -71% | 1 | 1 | 0% | 1,066 | 954 | -11% | 0 | 0 | — |
case-14 | pass→pass | 8,642 | 2,035 | -76% | 1 | 1 | 0% | 1,252 | 930 | -26% | 0 | 0 | — |
case-15 | pass→pass | 7,199 | 1,487 | -79% | 1 | 1 | 0% | 1,045 | 778 | -26% | 0 | 0 | — |
case-16 | fail→pass | 7,553 | 2,105 | -72% | 1 | 1 | 0% | 1,088 | 892 | -18% | 0 | 0 | — |
case-17 | fail→pass | 7,312 | 1,989 | -73% | 1 | 1 | 0% | 1,017 | 903 | -11% | 0 | 0 | — |
case-18 | fail→fail | 6,519 | 2,065 | -68% | 1 | 1 | 0% | 1,014 | 865 | -15% | 0 | 0 | — |
case-19 | fail→pass | 11,123 | 2,007 | -82% | 1 | 1 | 0% | 1,676 | 866 | -48% | 0 | 0 | — |
case-20 | pass→pass | 12,062 | 10,328 | -14% | 1 | 1 | 0% | 1,998 | 2,283 | +14% | 0 | 0 | — |
case-21 | pass→pass | 11,151 | 8,839 | -21% | 1 | 1 | 0% | 2,002 | 2,112 | +5% | 0 | 0 | — |
case-22 | pass→pass | 4,935 | 4,133 | -16% | 1 | 1 | 0% | 834 | 1,287 | +54% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 17 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 17 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.