Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Read, inspect, and add local dev diary entries after diary setup. Use when the user asks to view today's diary, edit diary notes, add work to the diary, or check what has been logged.
.claude/skills/tetherto-qv-dev-diary/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -62% | 0% |
Operate on the local diary after setup. This is not the setup skill; use qv-dev-diary-setup for initialization.
Diary files live under ~/.config/qvac-pr-skills/diary/ and are never committed.
Status:
bashnode .cursor/skills/_lib/developer-workflow/diary-cli.mjs --status
Read today's diary:
bashnode .cursor/skills/_lib/developer-workflow/diary-cli.mjs --read
Read a specific date:
bashnode .cursor/skills/_lib/developer-workflow/diary-cli.mjs --read YYYY-MM-DD
Print the diary file path:
bashnode .cursor/skills/_lib/developer-workflow/diary-cli.mjs --path YYYY-MM-DD
Append an entry:
bashnode .cursor/skills/_lib/developer-workflow/diary-cli.mjs --append '{"title":"Reviewed PR #123","summary":"Approved after checking tests.","type":"pr-review","pr":"123","ticket":"TICKET-123","status":"done"}'
To edit existing diary text:
--path.ReadFile.Do not put diary contents in repo files. Do not commit diary files.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 2,326 | 4,226 | +82% | 1 | 1 | 0% | 283 | 550 | +94% | 0 | 0 | — |
case-02 | fail→fail | 4,425 | 6,997 | +58% | 1 | 1 | 0% | 710 | 752 | +6% | 0 | 0 | — |
case-03 | fail→fail | 2,982 | 4,983 | +67% | 1 | 1 | 0% | 369 | 687 | +86% | 0 | 0 | — |
case-04 | fail→pass | 10,568 | 1,984 | -81% | 1 | 1 | 0% | 1,774 | 606 | -66% | 0 | 0 | — |
case-05 | fail→pass | 6,575 | 2,140 | -67% | 1 | 1 | 0% | 1,196 | 793 | -34% | 0 | 0 | — |
case-06 | fail→fail | 4,714 | 6,018 | +28% | 1 | 1 | 0% | 246 | 750 | +205% | 0 | 0 | — |
case-07 | fail→pass | 4,928 | 1,683 | -66% | 1 | 1 | 0% | 777 | 664 | -15% | 0 | 0 | — |
case-08 | fail→pass | 9,110 | 2,201 | -76% | 1 | 1 | 0% | 1,512 | 626 | -59% | 0 | 0 | — |
case-09 | fail→fail | 8,837 | 1,390 | -84% | 1 | 1 | 0% | 1,316 | 500 | -62% | 0 | 0 | — |
case-10 | fail→pass | 11,437 | 2,231 | -80% | 1 | 1 | 0% | 1,850 | 701 | -62% | 0 | 0 | — |
case-11 | fail→pass | 10,909 | 1,039 | -90% | 1 | 1 | 0% | 1,571 | 479 | -70% | 0 | 0 | — |
case-12 | fail→fail | 3,663 | 5,822 | +59% | 1 | 1 | 0% | 582 | 626 | +8% | 0 | 0 | — |
case-13 | fail→pass | 9,161 | 3,442 | -62% | 1 | 1 | 0% | 1,410 | 893 | -37% | 0 | 0 | — |
case-19 | fail→pass | 7,754 | 1,592 | -79% | 1 | 1 | 0% | 1,210 | 553 | -54% | 0 | 0 | — |
case-14 | fail→fail | 6,522 | 5,735 | -12% | 1 | 1 | 0% | 1,053 | 698 | -34% | 0 | 0 | — |
case-15 | fail→pass | 5,226 | 2,005 | -62% | 1 | 1 | 0% | 674 | 639 | -5% | 0 | 0 | — |
case-16 | fail→pass | 11,155 | 2,196 | -80% | 1 | 1 | 0% | 1,931 | 708 | -63% | 0 | 0 | — |
case-17 | fail→pass | 3,973 | 2,166 | -45% | 1 | 1 | 0% | 654 | 714 | +9% | 0 | 0 | — |
case-18 | pass→pass | 3,783 | 2,982 | -21% | 1 | 1 | 0% | 605 | 806 | +33% | 0 | 0 | — |
case-20 | fail→pass | 10,698 | 2,218 | -79% | 1 | 1 | 0% | 2,049 | 733 | -64% | 0 | 0 | — |
case-21 | pass→pass | 9,735 | 9,982 | +3% | 1 | 1 | 0% | 1,649 | 1,588 | -4% | 0 | 0 | — |
case-22 | pass→pass | 6,622 | 3,853 | -42% | 1 | 1 | 0% | 1,143 | 955 | -16% | 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 16 counted toward the lift figure. The other 6 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 +55 percentage points is the difference between those two pass rates over the 16 comparable cases.
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.