Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →This skill should be used when the user asks to "validate a plugin", "optimize plugin", "check plugin quality", "review plugin structure", or "run plugin optimizer".
.claude/skills/fradser-optimize-plugin/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -83% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 35% | 0% |
Execute plugin validation and optimization workflow. Target: $ARGUMENTS
Load plugin-optimizer:plugin-best-practices skill using the Skill tool for component templates, tool invocation rules, and type classification.
Goal: Validate structure and detect issues. Orchestrator MUST NOT apply fixes.
Actions:
realpath and verify existence.claude-plugin/plugin.json existscommands/, agents/, skills/, hooks/${CLAUDE_PLUGIN_ROOT}/examples/ templatescommands/ exists with .md files, use AskUserQuestion tool to ask about migrating to skills structurepython3 ${CLAUDE_PLUGIN_ROOT}/scripts/validate-plugin.py "$TARGET"--check=structure,manifest,frontmatter,tools,tokens--json-v, --verboseGoal: Launch agent to apply ALL fixes. Orchestrator does NOT make fixes directly.
Condition: Always execute.
Actions:
plugin-optimizer:plugin-optimizer agent with the following prompt content:AskUserQuestion tool before applying template fixes, presenting violations with specific examples and before/after comparison).claude-plugin/plugin.json after fixes:Path Reference Rules:
./reference.md)${CLAUDE_PLUGIN_ROOT} paths${CLAUDE_PLUGIN_ROOT}/examples/Redundancy & Efficiency:
Goal: Verify fixes, generate report, and update documentation.
Actions:
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/validate-plugin.py "$TARGET"README.md to reflect current state (metadata, directory structure, usage instructions; do not append version history log)markdown## Plugin Validation Report ### Plugin: [name] Location: [absolute-path] Version: [old] -> [new] ### Summary [2-3 sentences with key statistics] ### Phase 1: Issues Detected #### Critical ([count]) - `file/path` - [Issue description] #### Warnings ([count]) - `file/path` - [Issue description] ### Phase 2: Fixes Applied #### Structure Fixes - [Fix description] #### Template Conformance - **Agents**: [Count] validated, [count] fixed - **Instruction-type Skills**: [Count] validated, [count] fixed - **Knowledge-type Skills**: [Count] validated, [count] fixed #### Redundancy Fixes - [Consolidations applied] ### Phase 3: Verification Results - Structure validation: [PASS/FAIL] - Manifest validation: [PASS/FAIL] - Component validation: [PASS/FAIL] - Tool patterns validation: [PASS/FAIL] - Token budgets validation: [PASS/FAIL] ### Token Budget Analysis - Skills analyzed: [count] - Tier 1 (Metadata ~100): [OK count], [WARNING count] - Tier 2 (SKILL.md under 5k): [OK count], [WARNING count], [CRITICAL count] - Tier 3 (References, effectively unlimited): [total tokens] ### Component Inventory - Commands: [count] found, [count] valid - Agents: [count] found, [count] valid - Skills: [count] found, [count] valid ### Remaining Issues [Issues that couldn't be auto-fixed with explanations] ### Overall Assessment [PASS/FAIL] - [Detailed reasoning]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,064 | 26,083 | +330% | 1 | 1 | 0% | 273 | 1,695 | +521% | 0 | 0 | — |
case-02 | fail→fail | 6,149 | 6,614 | +8% | 1 | 1 | 0% | 302 | 1,658 | +449% | 0 | 0 | — |
case-03 | fail→fail | 4,792 | 7,676 | +60% | 1 | 1 | 0% | 264 | 1,594 | +504% | 0 | 0 | — |
case-04 | pass→fail | 11,111 | 5,229 | -53% | 1 | 1 | 0% | 1,850 | 1,568 | -15% | 0 | 0 | — |
case-05 | pass→fail | 11,760 | 18,821 | +60% | 1 | 1 | 0% | 1,856 | 1,395 | -25% | 0 | 0 | — |
case-06 | pass→fail | 15,782 | 8,994 | -43% | 1 | 1 | 0% | 2,900 | 1,586 | -45% | 0 | 0 | — |
case-07 | pass→pass | 12,657 | 4,692 | -63% | 1 | 1 | 0% | 2,058 | 1,959 | -5% | 0 | 0 | — |
case-17 | pass→pass | 7,843 | 3,132 | -60% | 1 | 1 | 0% | 1,339 | 1,600 | +19% | 0 | 0 | — |
case-08 | fail→pass | 28,255 | 3,349 | -88% | 1 | 1 | 0% | 1,436 | 1,682 | +17% | 0 | 0 | — |
case-09 | fail→pass | 13,346 | 6,947 | -48% | 1 | 1 | 0% | 2,340 | 1,928 | -18% | 0 | 0 | — |
case-10 | pass→pass | 7,450 | 4,704 | -37% | 1 | 1 | 0% | 949 | 1,803 | +90% | 0 | 0 | — |
case-11 | fail→fail | 8,716 | 5,091 | -42% | 1 | 1 | 0% | 1,349 | 1,823 | +35% | 0 | 0 | — |
case-12 | fail→pass | 12,620 | 4,287 | -66% | 1 | 1 | 0% | 2,106 | 1,829 | -13% | 0 | 0 | — |
case-13 | pass→pass | 2,653 | 3,528 | +33% | 1 | 1 | 0% | 483 | 1,672 | +246% | 0 | 0 | — |
case-14 | pass→pass | 3,779 | 2,631 | -30% | 1 | 1 | 0% | 636 | 1,616 | +154% | 0 | 0 | — |
case-15 | pass→pass | 3,885 | 2,995 | -23% | 1 | 1 | 0% | 598 | 1,614 | +170% | 0 | 0 | — |
case-16 | fail→fail | 9,422 | 1,984 | -79% | 1 | 1 | 0% | 1,565 | 1,407 | -10% | 0 | 0 | — |
case-18 | fail→pass | 49,416 | 1,803 | -96% | 1 | 1 | 0% | 8,212 | 1,380 | -83% | 0 | 0 | — |
case-19 | fail→pass | 7,219 | 3,308 | -54% | 1 | 1 | 0% | 1,263 | 1,704 | +35% | 0 | 0 | — |
case-20 | fail→fail | 12,514 | 8,835 | -29% | 1 | 1 | 0% | 2,293 | 1,755 | -23% | 0 | 0 | — |
case-21 | fail→pass | 9,307 | 4,211 | -55% | 1 | 1 | 0% | 1,572 | 1,745 | +11% | 0 | 0 | — |
case-22 | fail→pass | 13,296 | 12,976 | -2% | 1 | 1 | 0% | 2,032 | 2,845 | +40% | 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 15 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 +18 percentage points is the difference between those two pass rates over the 15 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.