Install any skill in seconds. Free to start, no credit card required.
Get Started Free →SpecKit Safe Update
.claude/skills/lingxling-speckit-updater/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | 44% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -15% | 0% |
This skill provides safe update capabilities for GitHub SpecKit installations, preserving customizations while applying template updates.
Installation: Available via plugin (/plugin marketplace add NotMyself/claude-plugins then /plugin install speckit-updater) or manual Git clone. See README.md for details.
When the user invokes /speckit-updater, you should:
powershell pwsh -NoProfile -Command "& 'C:\Users\bobby\.claude\skills\speckit-updater\scripts\update-wrapper.ps1'"
[PROMPT_FOR_APPROVAL] - Update scenario (existing SpecKit installation)[PROMPT_FOR_INSTALL] - Fresh installation scenario (no .specify/ directory)[PROMPT_FOR_APPROVAL] marker found):-Proceed flag[PROMPT_FOR_INSTALL] marker found):-Proceed flag to the user (this is an implementation detail)-Proceed flag-Proceed flag:powershell pwsh -NoProfile -Command "& 'C:\Users\bobby\.claude\skills\speckit-updater\scripts\update-wrapper.ps1' -Proceed"
Special cases:
-CheckOnly: run with that flag and show the report-Rollback: run with that flag and confirm restoration-Version: include that parameterUpdates SpecKit templates, commands, and scripts while preserving customizations.
Usage:
/speckit-updater - Interactive update/install with conversational approval workflow (recommended for Claude Code)/speckit-updater -Proceed - Proceed with update/install after approval (used by Claude after user confirms)/speckit-updater -CheckOnly - Check for updates without applying/speckit-updater -Version v0.0.72 - Update to specific version/speckit-updater -Force - Force overwrite SpecKit files (preserves custom commands)/speckit-updater -Rollback - Restore from previous backup/speckit-updater -Auto - DEPRECATED: Use conversational workflow instead (shows warning, maps to -Proceed)Fresh Installation (No .specify/ directory):
[PROMPT_FOR_INSTALL] marker-Proceed flag automatically (implementation detail hidden from user).specify/ structure, downloads templates, creates manifestProcess:
When you invoke this command, I will:
[PROMPT_FOR_APPROVAL] for updates, [PROMPT_FOR_INSTALL] for fresh installations)-Proceed flag)-Proceed flag to executeConversational Workflow: The skill uses a two-step approval process:
-Proceed → applies updatesRequirements:
The script is located at: {skill_path}/scripts/update-wrapper.ps1 (entry point) and {skill_path}/scripts/update-orchestrator.ps1 (main logic)
Entry point command:
powershellpwsh -NoProfile -Command "& '{skill_path}/scripts/update-wrapper.ps1' [parameters]"
Note: Both PowerShell-style (-CheckOnly) and Linux-style (--check-only) flags are supported via the wrapper script.
.specify/manifest.json with file hashes, version info, and backup history.specify/backups/ with automatic retention management--check-only shows exactly what would change without applying updates/speckit.constitution)--force| Code | Meaning | |------|---------| | 0 | Success | | 1 | General error | | 2 | Prerequisites not met | | 3 | Network/API error | | 4 | Git error | | 5 | User cancelled | | 6 | Rollback required (automatic) |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-14 | fail→pass | 10,746 | 2,500 | -77% | 1 | 1 | 0% | 1,704 | 2,453 | +44% | 0 | 0 | — |
case-01 | fail→fail | 1,920 | 6,249 | +225% | 1 | 1 | 0% | 233 | 2,429 | +942% | 0 | 0 | — |
case-02 | fail→fail | 3,302 | 9,408 | +185% | 1 | 1 | 0% | 313 | 2,400 | +667% | 0 | 0 | — |
case-03 | fail→fail | 7,198 | 7,599 | +6% | 1 | 1 | 0% | 907 | 2,370 | +161% | 0 | 0 | — |
case-04 | fail→fail | 11,047 | 6,422 | -42% | 1 | 1 | 0% | 331 | 2,528 | +664% | 0 | 0 | — |
case-05 | fail→fail | 5,778 | 6,941 | +20% | 1 | 1 | 0% | 814 | 2,314 | +184% | 0 | 0 | — |
case-06 | fail→fail | 9,536 | 7,088 | -26% | 1 | 1 | 0% | 1,016 | 2,437 | +140% | 0 | 0 | — |
case-07 | fail→fail | 13,499 | 9,614 | -29% | 1 | 1 | 0% | 1,724 | 2,548 | +48% | 0 | 0 | — |
case-12 | pass→pass | 6,632 | 2,136 | -68% | 1 | 1 | 0% | 1,062 | 2,368 | +123% | 0 | 0 | — |
case-08 | pass→pass | 13,896 | 4,369 | -69% | 1 | 1 | 0% | 1,682 | 2,777 | +65% | 0 | 0 | — |
case-09 | pass→pass | 8,835 | 3,656 | -59% | 1 | 1 | 0% | 1,109 | 2,577 | +132% | 0 | 0 | — |
case-10 | pass→pass | 10,361 | 4,132 | -60% | 1 | 1 | 0% | 1,602 | 2,657 | +66% | 0 | 0 | — |
case-11 | fail→pass | 11,207 | 2,006 | -82% | 1 | 1 | 0% | 1,947 | 2,366 | +22% | 0 | 0 | — |
case-13 | fail→pass | 25,916 | 2,435 | -91% | 1 | 1 | 0% | 2,910 | 2,433 | -16% | 0 | 0 | — |
case-15 | fail→pass | 10,572 | 1,697 | -84% | 1 | 1 | 0% | 1,828 | 2,391 | +31% | 0 | 0 | — |
case-16 | fail→pass | 16,329 | 3,009 | -82% | 1 | 1 | 0% | 2,714 | 2,306 | -15% | 0 | 0 | — |
case-17 | fail→pass | 15,633 | 1,753 | -89% | 1 | 1 | 0% | 2,416 | 2,367 | -2% | 0 | 0 | — |
case-18 | fail→pass | 11,806 | 4,205 | -64% | 1 | 1 | 0% | 1,595 | 2,548 | +60% | 0 | 0 | — |
case-19 | fail→pass | 15,125 | 5,939 | -61% | 1 | 1 | 0% | 1,849 | 2,908 | +57% | 0 | 0 | — |
case-20 | pass→pass | 16,691 | 13,993 | -16% | 1 | 1 | 0% | 2,542 | 4,174 | +64% | 0 | 0 | — |
case-21 | pass→pass | 6,726 | 11,503 | +71% | 1 | 1 | 0% | 1,083 | 3,860 | +256% | 0 | 0 | — |
case-22 | pass→pass | 25,149 | 7,401 | -71% | 1 | 1 | 0% | 3,452 | 2,467 | -29% | 0 | 0 | — |
case-23 | fail→pass | 14,320 | 11,049 | -23% | 1 | 1 | 0% | 2,021 | 3,952 | +96% | 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. 23 cases were attempted, and 16 counted toward the lift figure. The other 7 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 +39 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.