Compose Agents provides three orchestration primitives (SequentialAgent, ParallelAgent, LoopAgent) for building robust multi-agent pipelines. Use SequentialAgent to chain agents so each agent's output becomes the next agent's input; ParallelAgent to run analyses or tasks concurrently and gather results; LoopAgent to repeat sub-agents until a stop condition or approval is reached. Concrete examples include research->writing pipelines, parallel market and technical analysis, and review loops that iterate until quality thresholds are met.
Use this skill when you need to coordinate multiple specialized agents to solve a larger task: synthesising research from multiple analysts, running independent checks in parallel, or implementing iterative review-and-refine loops (e.g., draft -> review -> revise). It's ideal for content pipelines, multi-perspective analysis, and workflows that require retries or gated approvals.
Likely Python-based LLM agents and frameworks that run local/remote LLMs. Examples: Python LlmAgent integrations, agents built with LangChain-style adapters, and CLI/server deployments.
This skill has not been reviewed by our automated audit pipeline yet.
