| name | aistack-orchestrate |
| description | Orchestrate a multi-agent run with aistack — pick the right specialized agent (coder, architect, reviewer, researcher, tester, security-auditor, etc.), spawn it via the aistack MCP server, and coordinate handoffs. Use when a task is large enough to benefit from decomposition across specialized agents, or when the user asks to "orchestrate", "spawn an agent", or "run a multi-agent workflow". |
aistack: Orchestrate a multi-agent run
aistack turns Claude Code into a multi-agent orchestrator. Instead of one generalist doing everything, you decompose work and route each part to a specialized agent backed by the bundled aistack MCP server.
When to use
- A task spans distinct skills (design + implement + test + review).
- You want a dedicated critic (adversarial / reviewer) separate from the implementer.
- The user explicitly asks to orchestrate, spawn agents, or run a workflow.
For a small, single-step change, do not orchestrate — just do it directly.
Available agent types
| Type | Best at | Default model |
|---|
coder | Implement / refactor / debug | sonnet |
architect | System design, trade-offs | opus |
reviewer | Code review against conventions | sonnet |
adversarial | Break the change, find defects | opus |
tester | Write and run tests | sonnet |
researcher | Investigate, gather sources | sonnet |
security-auditor | OWASP-style security review | opus |
analyst | Data / metrics analysis | sonnet |
devops | CI/CD, infra, containers | sonnet |
documentation | Docs and references | sonnet |
coordinator | Plan and route across agents | opus |
grader | Score outcomes against a rubric | sonnet |
How to orchestrate
- Decompose: break the task into 1–4 sub-tasks, each mapped to one agent type above.
- Discover tools: the
aistack MCP server exposes agent_types, agent_spawn, agent_list, agent_status, agent_stop, task_create, task_assign. List them with npx @blackms/aistack mcp tools if unsure.
- Spawn: call
agent_spawn with the chosen type and a focused prompt (include acceptance criteria and the relevant file paths). Spawn background where possible so you can fan out.
- Coordinate: track progress with
agent_list / agent_status. Pass artifacts between agents via shared memory (see the aistack-memory skill) rather than re-deriving them.
- Converge: when sub-tasks finish, have a
reviewer or adversarial agent validate the combined result before reporting back.
Tips
- Keep each agent's prompt narrow — one responsibility. Broad prompts dilute the benefit of specialization.
- Prefer
opus-backed agents (architect, adversarial, security-auditor, coordinator) for judgment-heavy steps; sonnet for execution.
- Export agents as native Claude Code subagents with
npx @blackms/aistack export-agents if you want to @-mention them directly.
Reference: https://github.com/blackms/aistack
