[IMPORTANT] Use TaskCreate to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip.
Quick Summary
Goal: Optimize an existing skill for token efficiency, clarity, and effectiveness.
Workflow:
- Analyze — Review skill structure, line count, progressive disclosure
- Optimize — Reduce SKILL.md size, move details to references, improve clarity
- Validate — Verify skill still works correctly after optimization
Key Rules:
- Delegates to
skill-creator for optimization patterns
- SKILL.md target: under 100 lines with progressive disclosure
- Reference files also under 100 lines each
Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).
Think harder.
Use skill-creator and claude-code skills.
Use docs-seeker skills to search for documentation if needed.
Arguments
SKILL: $1 (default: *)
PROMPT: $2 (default: empty)
Your mission
Optimize an existing skill in .claude/skills/${SKILL} directory.
Mode detection:
- If arguments contain "auto" or "trust me": Skip plan approval, implement directly.
- Otherwise: Propose plan first, ask user to review before implementing.
Additional instructions
$PROMPT
Output Requirements
An output implementation plan must also follow the progressive disclosure structure:
- Always keep in mind that
SKILL.md and reference files should be token consumption efficient, so that progressive disclosure can be leveraged at best.
SKILL.md is always short and concise, straight to the point, treat it as a quick reference guide.- Create a directory using naming pattern from
## Naming section.
- Save the overview access point at
plan.md, keep it generic, under 80 lines, and list each phase with status/progress and links.
- For each phase, add
phase-XX-phase-name.md files containing sections (Context links, Overview with date/priority/statuses, Key Insights, Requirements, Architecture, Related code files, Implementation Steps, Todo list, Success Criteria, Risk Assessment, Security Considerations, Next steps).
IMPORTANT:
- Skills are not documentation, they are practical instructions for Claude Code to use the tools, packages, plugins or APIs to achieve the tasks.
- Each skill teaches Claude how to perform a specific development task, not what a tool does.
- Claude Code can activate multiple skills automatically to achieve the user's request.
IMPORTANT Task Planning Notes (MUST FOLLOW)
- Always plan and break work into many small todo tasks
- Always add a final review todo task to verify work quality and identify fixes/enhancements
1---2name: skill-optimize3description: [Skill Management] Optimize an existing agent skill4---5
6> **[IMPORTANT]** Use `TaskCreate` to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip.
7
8## Quick Summary
9
10**Goal:** Optimize an existing skill for token efficiency, clarity, and effectiveness.
11
12**Workflow:**
13
141. **Analyze** — Review skill structure, line count, progressive disclosure
152. **Optimize** — Reduce SKILL.md size, move details to references, improve clarity
163. **Validate** — Verify skill still works correctly after optimization
17
18**Key Rules:**
19
20- Delegates to `skill-creator` for optimization patterns
21- SKILL.md target: under 100 lines with progressive disclosure
22- Reference files also under 100 lines each
23
24**Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).**
25
26Think harder.
27Use `skill-creator` and `claude-code` skills.
28Use `docs-seeker` skills to search for documentation if needed.
29
30## Arguments
31
32SKILL: $1 (default: `*`)
33PROMPT: $2 (default: empty)
34
35## Your mission
36
37Optimize an existing skill in `.claude/skills/${SKILL}` directory.
38
39**Mode detection:**
40
41- If arguments contain "auto" or "trust me": Skip plan approval, implement directly.
42- Otherwise: Propose plan first, ask user to review before implementing.
43
44## Additional instructions
45
46<additional-instructions>$PROMPT</additional-instructions>
47
48## Output Requirements
49
50An output implementation plan must also follow the progressive disclosure structure:
51
52- Always keep in mind that `SKILL.md` and reference files should be token consumption efficient, so that **progressive disclosure** can be leveraged at best.
53- `SKILL.md` is always short and concise, straight to the point, treat it as a quick reference guide.
54- Create a directory using naming pattern from `## Naming` section.
55- Save the overview access point at `plan.md`, keep it generic, under 80 lines, and list each phase with status/progress and links.
56- For each phase, add `phase-XX-phase-name.md` files containing sections (Context links, Overview with date/priority/statuses, Key Insights, Requirements, Architecture, Related code files, Implementation Steps, Todo list, Success Criteria, Risk Assessment, Security Considerations, Next steps).
57
58**IMPORTANT:**
59
60- Skills are not documentation, they are practical instructions for Claude Code to use the tools, packages, plugins or APIs to achieve the tasks.
61- Each skill teaches Claude how to perform a specific development task, not what a tool does.
62- Claude Code can activate multiple skills automatically to achieve the user's request.
63
64---
65
66**IMPORTANT Task Planning Notes (MUST FOLLOW)**
67
68- Always plan and break work into many small todo tasks
69- Always add a final review todo task to verify work quality and identify fixes/enhancements
majiayu000