Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate Mos Sparkle appcast.xml from the latest build zip and recent git changes (since a given commit), then sync to docs/ for publishing.
.claude/skills/aiskillstore-generate-sparkle-appcast/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -67% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -62% | 0% |
Use this skill when the user wants to publish a new Mos release (stable or beta) and needs:
appcast.xml generated from the notarized .zip in build/zh* locales (Simplified/Traditional/HK/TW), and English for everything elseInputs
--since <commit>: the previous release commit (exclusive). Used to generate release notes from changes since that commit.build/ named:Mos.Versions.<version>-<YYYYMMDD>.<num>.zip (stable)Mos.Versions.<version>-beta-<YYYYMMDD>.<num>.zip (beta)sparkle_private_key.txt (gitignored).RELEASE_NOTES_BASE_URL (default https://mos.caldis.me/release-notes)RELEASE_NOTES_ZH_FILE / RELEASE_NOTES_EN_FILE to point to pre-written HTML files (otherwise the script writes to build/release-notes/<tag>.*.html)What to do
bash .codex/skills/generate-sparkle-appcast/scripts/generate_appcast.sh --since <commit>build/appcast.xml (generated)docs/appcast.xml (copied for mos.caldis.me/appcast.xml)build/release-notes/<tag>.zh.html + build/release-notes/<tag>.en.html (generated)docs/release-notes/<tag>.zh.html + docs/release-notes/<tag>.en.html (copied for hosting)Notes
<sparkle:releaseNotesLink> entries: xml:lang="zh" points to the Chinese page, and the default link points to the English page.build/release-notes/<tag>.zh.html and build/release-notes/<tag>.en.html before running the script; the script will reuse them if present, otherwise it generates a default template from git history.sparkle:edSignature.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-08 | pass→pass | 15,765 | 11,626 | -26% | 1 | 1 | 0% | 1,995 | 1,673 | -16% | 0 | 0 | — |
case-09 | fail→pass | 36,487 | 2,720 | -93% | 1 | 1 | 0% | 1,789 | 1,000 | -44% | 0 | 0 | — |
case-01 | fail→fail | 22,124 | 15,725 | -29% | 1 | 1 | 0% | 3,185 | 1,051 | -67% | 0 | 0 | — |
case-02 | fail→fail | 12,618 | 11,039 | -13% | 1 | 1 | 0% | 2,271 | 861 | -62% | 0 | 0 | — |
case-03 | fail→fail | 16,642 | 16,705 | +0% | 1 | 1 | 0% | 1,921 | 852 | -56% | 0 | 0 | — |
case-04 | pass→pass | 15,381 | 11,495 | -25% | 1 | 1 | 0% | 2,333 | 1,767 | -24% | 0 | 0 | — |
case-05 | pass→pass | 14,965 | 10,660 | -29% | 1 | 1 | 0% | 1,614 | 1,536 | -5% | 0 | 0 | — |
case-06 | pass→pass | 22,427 | 18,176 | -19% | 1 | 1 | 0% | 2,827 | 3,225 | +14% | 0 | 0 | — |
case-07 | fail→pass | 25,248 | 9,000 | -64% | 1 | 1 | 0% | 3,327 | 1,305 | -61% | 0 | 0 | — |
case-10 | fail→pass | 23,613 | 8,743 | -63% | 1 | 1 | 0% | 3,924 | 1,276 | -67% | 0 | 0 | — |
case-11 | fail→pass | 15,957 | 7,585 | -52% | 1 | 1 | 0% | 3,005 | 1,129 | -62% | 0 | 0 | — |
case-12 | fail→pass | 55,834 | 6,760 | -88% | 1 | 1 | 0% | 2,330 | 891 | -62% | 0 | 0 | — |
case-13 | pass→pass | 17,683 | 9,041 | -49% | 1 | 1 | 0% | 1,738 | 1,132 | -35% | 0 | 0 | — |
case-14 | pass→pass | 12,739 | 2,414 | -81% | 1 | 1 | 0% | 1,427 | 1,012 | -29% | 0 | 0 | — |
case-15 | pass→pass | 7,607 | 8,692 | +14% | 1 | 1 | 0% | 1,189 | 1,298 | +9% | 0 | 0 | — |
case-16 | fail→pass | 13,492 | 9,216 | -32% | 1 | 1 | 0% | 2,264 | 1,325 | -41% | 0 | 0 | — |
case-17 | fail→pass | 13,190 | 8,421 | -36% | 1 | 1 | 0% | 1,834 | 1,226 | -33% | 0 | 0 | — |
case-18 | fail→pass | 17,187 | 7,916 | -54% | 1 | 1 | 0% | 1,912 | 1,073 | -44% | 0 | 0 | — |
case-19 | pass→pass | 6,432 | 8,623 | +34% | 1 | 1 | 0% | 1,087 | 1,228 | +13% | 0 | 0 | — |
case-20 | fail→pass | 13,735 | 7,952 | -42% | 1 | 1 | 0% | 1,393 | 943 | -32% | 0 | 0 | — |
case-21 | fail→pass | 21,621 | 7,063 | -67% | 1 | 1 | 0% | 2,708 | 1,005 | -63% | 0 | 0 | — |
case-22 | fail→pass | 26,316 | 7,032 | -73% | 1 | 1 | 0% | 3,663 | 810 | -78% | 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 18 counted toward the lift figure. The other 4 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 +50 percentage points is the difference between those two pass rates over the 18 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.