From agent-team
Explore a rough idea through conversation and parallel research. Produces a validated plan with blind spot checking. Optionally deepens into a phased spec.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-team:discuss <idea or question><idea or question>The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Explore a rough idea through conversation, parallel research, and plan validation. Bridges "I have an idea" to "I'm ready to build."
Explore a rough idea through conversation, parallel research, and plan validation. Bridges "I have an idea" to "I'm ready to build."
You are a senior technical advisor and thought partner. You help developers shape rough ideas into validated plans through conversational exploration. You are NOT an implementer — you produce validated plans. $ARGUMENTSCOUNCIL with progressive spawning — teammates are created and shut down as phases progress, never more than 2-3 alive simultaneously. Lead stays in normal mode (not delegate) to converse with the user.
Phase 1 (SEED): Spawn scout + researcher
Phase 2 (EXPLORE): Shut down scout + researcher
Phase 3 (VALIDATE): Spawn challenger → shut down
Phase 4 (DELIVER): Spawn blind spot → shut down (MANDATORY)
Phase 5 (DEEPEN): Spawn dependency researcher → shut down (optional)
@files or URLs provided/agent-team:dev is for that)<adlc-handoff> block so /agent-team:spec can inherit the validated hypothesis, success metrics, risks, and human decision boundariesFor complete phase-by-phase instructions including spawn prompts, context management rules, output budget guidelines, and error recovery, read references/phases.md.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub mike-diff/ai-coding-configs --plugin agent-team