Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →GitHub CLI (gh) command reference. Use when working with GitHub repositories, PRs, issues, actions, `gh api`, or any GitHub operations from the command line.
.claude/skills/fredrikaverpil-gh-cli/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-11 | ✓→✗ | ▼ Worse | -30% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 67% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 105% | 0% |
gh --help and https://cli.github.com/manual cover the command surface. This skill covers only what they don't.
--limit N on every list command (pr list, issue list, run list).The defaults dump far more than a terminal needs.
--json FIELDS when the output feeds a script; pipe to jq from there.gh pr review only submits an overall body. Inline comments need gh api, which posts the comments and the summary in one call. Read the diff first (gh pr diff NUMBER), then:
bashgh api repos/{owner}/{repo}/pulls/NUMBER/reviews --input - <<'EOF' { "commit_id": "LATEST_COMMIT_SHA", "event": "COMMENT", "body": "Overall: solid changes with a few suggestions.", "comments": [ { "path": "src/example.go", "line": 42, "side": "RIGHT", "body": "This variable is unused." } ] } EOF
side is RIGHT for additions, LEFT for deletions; multi-line comments add start_line and start_side. A body may contain a GitHub suggestion block ( suggestion ) to propose replacement code.
event is COMMENT, APPROVE, or REQUEST_CHANGES — ask me which one before submitting, and default to COMMENT.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 11,886 | 9,075 | -24% | 1 | 1 | 0% | 2,117 | 1,679 | -21% | 0 | 0 | — |
case-02 | pass→pass | 2,182 | 1,386 | -36% | 1 | 1 | 0% | 378 | 632 | +67% | 0 | 0 | — |
case-03 | pass→pass | 1,963 | 1,715 | -13% | 1 | 1 | 0% | 319 | 654 | +105% | 0 | 0 | — |
case-04 | pass→pass | 2,167 | 1,620 | -25% | 1 | 1 | 0% | 306 | 585 | +91% | 0 | 0 | — |
case-05 | pass→pass | 3,750 | 4,217 | +12% | 1 | 1 | 0% | 522 | 788 | +51% | 0 | 0 | — |
case-06 | pass→pass | 7,925 | 3,484 | -56% | 1 | 1 | 0% | 1,391 | 932 | -33% | 0 | 0 | — |
case-07 | pass→pass | 7,494 | 1,932 | -74% | 1 | 1 | 0% | 1,174 | 649 | -45% | 0 | 0 | — |
case-08 | fail→pass | 13,077 | 1,944 | -85% | 1 | 1 | 0% | 2,186 | 678 | -69% | 0 | 0 | — |
case-09 | pass→pass | 5,824 | 2,433 | -58% | 1 | 1 | 0% | 1,066 | 679 | -36% | 0 | 0 | — |
case-10 | pass→pass | 4,739 | 1,587 | -67% | 1 | 1 | 0% | 743 | 593 | -20% | 0 | 0 | — |
case-11 | pass→fail | 4,423 | 1,596 | -64% | 1 | 1 | 0% | 902 | 628 | -30% | 0 | 0 | — |
case-12 | pass→pass | 4,693 | 1,454 | -69% | 1 | 1 | 0% | 829 | 623 | -25% | 0 | 0 | — |
case-13 | pass→pass | 6,160 | 2,267 | -63% | 1 | 1 | 0% | 1,098 | 691 | -37% | 0 | 0 | — |
case-14 | pass→pass | 7,836 | 3,785 | -52% | 1 | 1 | 0% | 1,399 | 1,097 | -22% | 0 | 0 | — |
case-15 | pass→pass | 6,112 | 1,720 | -72% | 1 | 1 | 0% | 1,024 | 615 | -40% | 0 | 0 | — |
case-16 | pass→pass | 3,656 | 2,074 | -43% | 1 | 1 | 0% | 592 | 696 | +18% | 0 | 0 | — |
case-17 | pass→pass | 3,544 | 1,424 | -60% | 1 | 1 | 0% | 681 | 571 | -16% | 0 | 0 | — |
case-18 | pass→pass | 3,767 | 1,968 | -48% | 1 | 1 | 0% | 639 | 712 | +11% | 0 | 0 | — |
case-19 | pass→pass | 4,264 | 1,336 | -69% | 1 | 1 | 0% | 821 | 602 | -27% | 0 | 0 | — |
case-20 | pass→pass | 2,986 | 1,153 | -61% | 1 | 1 | 0% | 509 | 562 | +10% | 0 | 0 | — |
case-21 | pass→pass | 2,796 | 1,546 | -45% | 1 | 1 | 0% | 452 | 583 | +29% | 0 | 0 | — |
case-22 | pass→pass | 2,712 | 1,648 | -39% | 1 | 1 | 0% | 502 | 589 | +17% | 0 | 0 | — |
case-23 | pass→pass | 1,811 | 1,440 | -20% | 1 | 1 | 0% | 266 | 583 | +119% | 0 | 0 | — |
case-24 | pass→pass | 2,437 | 3,111 | +28% | 1 | 1 | 0% | 448 | 738 | +65% | 0 | 0 | — |
case-25 | pass→pass | 1,578 | 1,213 | -23% | 1 | 1 | 0% | 261 | 527 | +102% | 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. 25 cases were attempted. The headline lift of +4 percentage points is the difference between those two pass rates over the 25 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.