/agents:analyze

Analyze the plugin collection to identify where sub-agents would improve workflows by isolating verbose output, enforcing constraints, or specializing behavior.

Context

• Plugin directories: !find . -maxdepth 1 -type d -name '*-plugin' 2>/dev/null
• Existing agents: !find agents-plugin/agents -maxdepth 1 -name '*.md' 2>/dev/null
• Skills: !find . -path '*/skills/*/skill.md' 2>/dev/null
• Skills (user-invocable): !find . -path '*/skills/*/SKILL.md' -not -path './agents-plugin/*' 2>/dev/null

Parameters

• $1: Optional --focus <plugin-name> to analyze a single plugin in depth

Your Task

Perform a systematic analysis of the plugin collection to identify sub-agent opportunities.

Step 1: Inventory Current State

Scan the repository to build an inventory:

1. List all plugins with their skill/command counts
2. Read existing agents in agents-plugin/agents/ to understand current coverage
3. If --focus is provided, restrict analysis to that plugin only

Step 2: Identify Sub-Agent Opportunities

For each plugin (or focused plugin), evaluate skills and commands against these criteria:

Context Isolation (Primary Value)

Operations that produce verbose output benefiting from isolation:

Constraint Enforcement

Operations that should be limited to specific tools:

Model Selection Opportunities

Step 3: Gap Analysis

Compare identified opportunities against existing agents:

1. Missing agents: Skills that have no corresponding agent
2. Model mismatches: Agents using wrong model for their task complexity
3. Tool over-permissions: Agents with tools they don’t need
4. Consolidation opportunities: Multiple agents that could be merged
5. Delegation mapping: Check if /delegate references agents that don’t exist

Step 4: Produce Recommendations

For each recommended new agent, specify:

### Proposed: <agent-name>

- **Model**: opus | haiku
- **Covers plugins**: <list>
- **Context value**: <what verbose output it isolates>
- **Tools**: <minimal set>
- **Constraint**: <read-only, no-network, etc.>
- **Priority**: HIGH | MEDIUM | LOW
- **Rationale**: <why this is better than inline execution>

For model/tool corrections to existing agents:

### Fix: <agent-name>

- **Current model**: X → **Recommended**: Y
- **Reason**: <why the change improves things>

Step 5: Implementation Check

If new agents are recommended, check:

• Agent name doesn’t conflict with existing
• Agent fills a gap referenced by /delegate command
• Model selection follows haiku-for-mechanical, opus-for-reasoning
• Tool set is minimal (principle of least privilege)
• Agent has clear “does / does NOT do” boundaries

Output Format

## Sub-Agent Analysis Report

**Scope**: [All plugins | focused plugin name]
**Date**: [today]
**Plugins analyzed**: N
**Existing agents**: N
**Skills without agent coverage**: N

### Current Coverage Map

| Domain | Agent | Skills Covered | Gaps |
|--------|-------|----------------|------|
| ... | ... | ... | ... |

### Recommended New Agents

[Proposals from Step 4]

### Recommended Fixes

[Model/tool corrections from Step 4]

### Delegation Mapping Updates

[Any updates needed for /delegate command's agent reference table]

### Priority Summary

| Priority | Count | Top Recommendation |
|----------|-------|-------------------|
| HIGH | N | ... |
| MEDIUM | N | ... |
| LOW | N | ... |

Post-Actions

After presenting the analysis:

1. Ask the user if they want to implement any of the recommendations
2. If yes, create the agent files following the existing patterns in agents-plugin/agents/
3. Update agents-plugin/README.md with new agents
4. Update /delegate command’s agent reference table if needed