Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Draft user-facing release notes, store changelogs, and internal publish summaries.
.claude/skills/hoangnguyen0403-publish-notes/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -4% | 0% |
> !IMPORTANT] > Draft user-facing release notes, store changelogs, and internal publish summaries.
Optional args: slug=<feature>, ticket=<id/url>, mode=interactive|autonomous|channel, channel=<id>, auto_continue=true|false, profile=business|hybrid|technical.
When the user asks to perform this workflow, execute the following steps:
Goal: Convert verified changes into accurate user-facing and internal release notes.
common-operator-profile; carry the inherited operator_profile from the Handoff Payload without re-inferring it.operator_profile=business gets Public Notes as the primary artifact, Internal Notes as an appendix.retro-learn.deploy-release or uat-signoff to draft user-facing and internal release communication.slug, operator_profile, public notes, internal notes, security/privacy notes, verification source, next workflow.md# Release Notes: [Version] ## Public Notes ## Internal Notes ## Security Or Privacy Notes ## Verification Source ## Outcome Report feature_status: implemented requirement_trace: BRD-OBJ-* -> REQ-* -> AC-* -> SRS-* -> evidence completed_evidence: []; missing_evidence: []; decision_needed: []; recommended_next_workflow: retro-learn ## Next Workflow retro-learn ## Cost Report Call `get_session_cost(workflow="publish-notes")` before final handoff.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | fail→pass | 11,885 | 4,072 | -66% | 1 | 1 | 0% | 2,077 | 1,336 | -36% | 0 | 0 | — |
case-01 | fail→pass | 14,092 | 7,604 | -46% | 1 | 1 | 0% | 2,543 | 1,946 | -23% | 0 | 0 | — |
case-02 | fail→pass | 17,255 | 10,595 | -39% | 1 | 1 | 0% | 3,046 | 2,445 | -20% | 0 | 0 | — |
case-03 | fail→fail | 16,260 | 9,384 | -42% | 1 | 1 | 0% | 2,716 | 2,197 | -19% | 0 | 0 | — |
case-04 | pass→pass | 6,648 | 4,268 | -36% | 1 | 1 | 0% | 1,087 | 1,346 | +24% | 0 | 0 | — |
case-05 | pass→fail | 13,041 | 7,614 | -42% | 1 | 1 | 0% | 2,034 | 1,813 | -11% | 0 | 0 | — |
case-06 | pass→fail | 6,739 | 7,183 | +7% | 1 | 1 | 0% | 1,464 | 1,914 | +31% | 0 | 0 | — |
case-07 | fail→pass | 9,165 | 7,920 | -14% | 1 | 1 | 0% | 1,573 | 1,979 | +26% | 0 | 0 | — |
case-08 | pass→pass | 13,446 | 7,364 | -45% | 1 | 1 | 0% | 2,170 | 1,825 | -16% | 0 | 0 | — |
case-09 | fail→pass | 7,370 | 3,816 | -48% | 1 | 1 | 0% | 1,221 | 1,169 | -4% | 0 | 0 | — |
case-10 | pass→pass | 10,764 | 6,464 | -40% | 1 | 1 | 0% | 1,592 | 1,726 | +8% | 0 | 0 | — |
case-11 | fail→pass | 9,013 | 2,366 | -74% | 1 | 1 | 0% | 1,533 | 915 | -40% | 0 | 0 | — |
case-12 | fail→pass | 4,561 | 6,446 | +41% | 1 | 1 | 0% | 750 | 1,799 | +140% | 0 | 0 | — |
case-13 | fail→pass | 7,717 | 5,221 | -32% | 1 | 1 | 0% | 1,406 | 1,527 | +9% | 0 | 0 | — |
case-19 | fail→pass | 10,153 | 3,724 | -63% | 1 | 1 | 0% | 1,989 | 1,240 | -38% | 0 | 0 | — |
case-14 | fail→pass | 10,208 | 5,854 | -43% | 1 | 1 | 0% | 2,002 | 1,607 | -20% | 0 | 0 | — |
case-15 | pass→pass | 10,404 | 11,702 | +12% | 1 | 1 | 0% | 1,790 | 2,687 | +50% | 0 | 0 | — |
case-16 | pass→pass | 8,289 | 6,342 | -23% | 1 | 1 | 0% | 1,350 | 1,801 | +33% | 0 | 0 | — |
case-17 | pass→pass | 9,629 | 6,936 | -28% | 1 | 1 | 0% | 1,732 | 1,664 | -4% | 0 | 0 | — |
case-20 | fail→pass | 11,096 | 2,958 | -73% | 1 | 1 | 0% | 1,719 | 1,099 | -36% | 0 | 0 | — |
case-21 | pass→pass | 5,988 | 5,841 | -2% | 1 | 1 | 0% | 1,019 | 1,589 | +56% | 0 | 0 | — |
case-22 | pass→pass | 5,326 | 3,614 | -32% | 1 | 1 | 0% | 889 | 1,260 | +42% | 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. The headline lift of +41 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are 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.