Official guides, runtime docs, and framework references for deciding when to split one agent into a coordinated team or scheduled background workflow.
public · 9 sources
A source-backed method for deciding when one agent should become a team, then designing roles, handoffs, context boundaries, evals, and human checkpoints.
Anthropic's practical taxonomy for workflows, agents, routing, parallelization, orchestrator-workers, evaluator loops, and tool design.
OpenClaw guide to multiple isolated agents, workspaces, state directories, channel-account bindings, and per-agent tool or sandbox policy.
Hermes reference for spawning child agents with isolated context, restricted toolsets, separate terminal sessions, timeouts, and summary-only returns.
Hermes cron guide for one-shot and recurring scheduled agent runs, skill-backed jobs, delivery targets, lifecycle actions, and no-agent script jobs.
OpenAI Agents SDK guide to agents-as-tools, handoffs, specialist delegation, prompt iteration, monitoring, and evals.
Pattern comparison for subagents, handoffs, skills, routers, custom workflows, context engineering, and cost tradeoffs.
Microsoft AutoGen docs for creating teams, selecting team presets, observing behavior, and knowing when to start with one agent.
CrewAI docs for crews, tasks, sequential or hierarchical processes, manager agents, memory, planning, callbacks, and logging.
Loading comments…