brewcode%3askills（brewcode 3askills 搜索）

skills Skill

Skill Management: List, improve, create skills with activation optimization.

Phase 1: Parse Arguments

Extract mode and target from $ARGUMENTS:

Smart Detection: If first argument is NOT a mode keyword (list, up, create), treat entire input as target for up mode.

Examples:

• /brewcode:skills → list
• /brewcode:skills list → list
• /brewcode:skills up commit → up, target=commit
• /brewcode:skills up ~/.claude/skills/ → up, target=folder (all skills)
• /brewcode:skills create "semantic code search" → create, target=prompt
• /brewcode:skills create./spec.md → create, target=spec file
• /brewcode:skills commit → up, target=commit (shorthand)
• /brewcode:skills brewcode/skills/setup → up, target=path (shorthand)
• /brewcode:skills ~/.claude/skills/ → up, target=folder (shorthand)

Mode: list

List all skills (global, project, plugin).

EXECUTE using Bash tool:

bash "scripts/list-skills.sh" && echo "✅ list" || echo "❌ list FAILED"

STOP if ❌ — verify skill base directory is resolved and scripts exist.

Mode: up

Improve skill(s) via skill-creator agent.

Step 1: Resolve Target

EXECUTE using Bash tool — resolve target:

TARGET="$ARGUMENTS"
# Remove "up " prefix if present (explicit mode), else use as-is (shorthand)
if [[ "$TARGET" == up\ * ]] || [[ "$TARGET" == "up" ]]; then
  TARGET="${TARGET#up }"
  TARGET="${TARGET#up}"
fi
TARGET="$(echo "$TARGET" | xargs)"  # trim

if [[ -z "$TARGET" ]]; then
  echo "❌ No target specified. Usage: /brewcode:skills up <name|path|folder>"
  echo "Shorthand: /brewcode:skills <name|path|folder>"
  exit 1
fi

# Check if it's a path
if [[ -d "$TARGET" ]]; then
  echo "TYPE: folder"
  echo "PATH: $TARGET"
  find "$TARGET" -name "SKILL.md" -type f 2>/dev/null | head -20
elif [[ -f "$TARGET" ]]; then
  echo "TYPE: file"
  echo "PATH: $TARGET"
elif [[ -f "$TARGET/SKILL.md" ]]; then
  echo "TYPE: skill-dir"
  echo "PATH: $TARGET/SKILL.md"
else
  # Search by name
  echo "TYPE: name"
  echo "NAME: $TARGET"
  # Search locations
  for loc in ~/.claude/skills .claude/skills; do
    if [[ -f "$loc/$TARGET/SKILL.md" ]]; then
      echo "FOUND: $loc/$TARGET/SKILL.md"
    fi
  done
fi

Step 2: Spawn skill-creator Agent(s)

Single skill: Spawn one agent.

Task tool:
  subagent_type: "brewcode:skill-creator"
  prompt: |
    Improve this skill's activation rate and quality.

    Skill path: {SKILL_PATH}

    Tasks:
    1. Read current SKILL.md
    2. Analyze description for trigger keywords
    3. Check body for imperative form, clear instructions
    4. Apply skill-creator best practices
    5. Update SKILL.md with improvements

    Focus on:
    - Description has "Use when:", "Trigger keywords:", "Triggers -"
    - No summary in description (only triggers!)
    - Third-person voice
    - Imperative form in body
    - <500 lines
  model: opus

Multiple skills (folder): Spawn agents in parallel (single message).

# For each SKILL.md found, spawn in ONE Task call block:
Task tool:
  subagent_type: "brewcode:skill-creator"
  prompt: "Improve skill: {SKILL_PATH_1}..."

Task tool:
  subagent_type: "brewcode:skill-creator"
  prompt: "Improve skill: {SKILL_PATH_2}..."

# etc.

CRITICAL: Spawn ALL agents in a single message for parallel execution.

Mode: create

Research topic, then create skill via skill-creator.

Step 1: Determine Input Type

Step 2: Clarify Invocation Type

Use AskUserQuestion before spawning any agents:

header: "Invocation"
question: "Who will invoke this skill?"
options:
  - label: "User only (slash command)"
    description: "Only via /skill-name — sets disable-model-invocation: true, simple description"
  - label: "LLM auto-detect"
    description: "Claude picks it up from context — full trigger keyword optimization"
  - label: "Both (default)"
    description: "User slash command + LLM auto-detection"

Save answer as INVOCATION_TYPE for use in Step 4.

Step 3: Parallel Research

Spawn two agents in parallel (single message):

Task tool:
  subagent_type: "Explore"
  prompt: |
    Research codebase for: {TOPIC}

    Find:
    - Related existing skills
    - Patterns and conventions
    - Similar implementations
    - Relevant file structures

    Output: Summary of findings for skill creation.

Task tool:
  subagent_type: "general-purpose"
  prompt: |
    Web research for: {TOPIC}

    Search for:
    - Best practices for this type of skill
    - Similar tools/implementations
    - Common patterns and conventions

    Use WebSearch and WebFetch tools.
    Output: Summary of external knowledge for skill creation.

Step 4: Create Skill

After research completes, spawn skill-creator:

Task tool:
  subagent_type: "brewcode:skill-creator"
  prompt: |
    Create new skill based on research.

    Topic: {TOPIC}
    Invocation type: {INVOCATION_TYPE}

    ## Codebase Research
    {EXPLORE_RESULTS}

    ## Web Research
    {WEB_RESULTS}

    ## Requirements
    - Location: .claude/skills/{skill-name}/ (project) or ~/.claude/skills/{skill-name}/ (global)
    - Follow skill-creator best practices
    - Optimized description for 84% activation
    - Include README.md
  model: opus

Output Format

# skills [{MODE}]

## Detection

| Field | Value |
|-------|-------|
| Arguments | `$ARGUMENTS` |
| Mode | `[detected mode]` |
| Target | `[target or none]` |

## Results

[Mode-specific output]

## Skills Summary

| Location | Count | Skills |
|----------|-------|--------|
| Global (~/.claude/skills/) | N | skill1, skill2 |
| Project (.claude/skills/) | N | skill3 |
| Plugins | N | plugin:skill1 |

## Next Steps

- [recommendations based on mode]