founder-agent

name	founder-agent
description	Use when the user wants to build a reusable creative-ops system for a product or company — brand system, voice guide, audience map, brief library, prompt library. Triggers when the user says "build a brand system", "set up creative ops", "founder agent", "/founder-agent", or describes a launch/positioning/brand-foundation task. Invokes the AgentSuite Founder agent via MCP.

Founder Agent (AgentSuite)

This skill invokes the Founder agent from the AgentSuite MCP server. It produces 26 reusable creative-ops artifacts in 30-90 seconds and pauses for human approval before promoting them to long-lived storage.

When to use

User wants any of:

• Brand system / brand book
• Voice and tone guide
• Audience map / persona doc
• Positioning document
• Reusable creative brief library
• Prompt library for marketing assets
• Any "I keep restating context every time I write copy" problem

When NOT to use

• One-off content (a single landing page, single email) — overkill
• Technical product spec — that's the Product agent (v0.3+)
• Competitive intel / vendor analysis — that's the CIO agent (v0.7+)

Steps

1. Confirm AgentSuite MCP is configured. Call agentsuite_list_agents first. If it returns {"enabled": ["founder"]}, proceed. If the call fails, tell the user to add the AgentSuite MCP entry to their .mcp.json (paste the snippet from ~/.claude/skills/founder-agent/mcp-snippet.json).

2. Gather inputs from the user. Ask:

   • Business goal (one sentence)
   • Slug for _kernel/ promotion (lowercase, hyphenated)
   • Path to inputs directory (optional — directory of READMEs, voice samples, screenshots)

3. Invoke founder_run. Pass the gathered fields plus agent_name="founder", role_domain="creative-ops", user_request=<the user's original ask>. Wait for the JSON-RPC response (30-90 s; do not poll).

4. Show the user the result. Display the primary_path (always brand-system.md) and the list of open_questions. Tell them to read brand-system.md and qa_report.md.

5. Wait for approval signal. When the user says "approved" or "ship it", call founder_approve with the same run_id and project_slug.

6. Confirm promotion. Echo the promoted_paths from the ApprovalResult so the user sees what landed in _kernel/.

Cost expectation

A typical run costs $0.20 – $1.00 against Claude Sonnet or GPT-4o. Hard cap is $5.00 per run. If HardCapExceeded is raised, suggest the user reduce input size or raise AGENTSUITE_COST_CAP_USD.

Failure modes

• NoProviderConfigured — tell user to set ANTHROPIC_API_KEY or OPENAI_API_KEY in their shell.
• ConsistencyCheckFailed — open consistency_report.json and explain the mismatch.
• extract stage produced invalid JSON — re-run, usually transient.

After approval

The Design agent (v0.2+) can read directly from _kernel/<project_slug>/. Until then, the artifacts are ready to feed into ChatGPT, Claude, Midjourney, etc., as context for any creative work.