| name | bmad-ml-research-party |
| description | Run multi-agent research discourse session. Use when the user requests to "start a research party" or "run a journal club". |
Follow the instructions in ./workflow.md.
Optional: Agent Teams (Claude Code)
When to use. If CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 is set at session start, assemble a team instead of dispatching serially. Nosh is team lead; the default roster is bmad-sova (lit survey), bmad-cypher (mechanistic interp), bmad-viper (empirical), bmad-killjoy (eval), bmad-fade (theory). The roster is user-configurable per session — swap in other Research or Build specialists as needed.
Fallback contract. If the env var is unset, this workflow runs as serial Task-tool invocations per ./workflow.md. Users who don't opt in see no behavior change. The installer never writes CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS; users export it from their shell.