list-agents

List Agents — Discovery and Inspection

Find and display existing agents, skills, and workflows across the workspace.

Process

Step 1: Scan for Agents

Search these locations for agent definitions:

# Primary location
~/.talon/agents/*/AGENT.yaml

# Standalone skills
~/.talon/skills/*/SKILL.md

# Standalone workflows
~/.talon/workflows/*.yaml

Step 2: Display Agent Summary

For each agent found, show:

Agent: <name>
  Description: <from AGENT.yaml>
  Model: <model>
  Backend: <backend or "default">
  Skills: <count> (<skill-1>, <skill-2>, ...)
  Workflows: <count> (<workflow-1>, ...)
  Identity: <first line of SOUL.md>

Step 3: Detailed Inspection (if requested)

If the user asks about a specific agent, show:

1. Full SOUL.md — identity and philosophy
2. AGENT.yaml — complete configuration
3. Skills list — each skill's name, description, and step count
4. Workflow diagram — text representation of state machine:

   [planning] → [implementation] → [review] → [done]
                       ↑                |
                       └── (rejected) ──┘
   

5. Launch command — how to run this agent via dispatch

Step 4: Gap Analysis (optional)

If the user asks "what's missing?" or wants to improve an agent:

• Check if SOUL.md exists and has substance (not just template placeholders)
• Check if skills have validation steps
• Check if workflows pass validation
• Check if AGENT.yaml references skills/workflows that actually exist
• Suggest improvements based on domain best practices

Output Format

Summary view (default):

Found 3 agents:

  analyst         Financial research agent (opus)
                  Skills: dcf-valuation, fundamental-analysis, sec-filing-reader
                  Workflows: research-pipeline

  sentinel        Security audit agent (sonnet, default)
                  Skills: dependency-scan, owasp-check, secret-detection

  scribe          Documentation agent (haiku, default)
                  Skills: api-doc-gen, changelog-writer