Agent orchestration patterns for agentic loops, multi-agent coordination, alternative frameworks, and multi-scenario workflows. Use when building autonomous agent loops, coordinating multiple agents, evaluating CrewAI/AutoGen/Swarm, or orchestrating complex multi-step scenarios.
Instrucciones de origen · Vista previa de solo lectura
name
agent-orchestration
license
MIT
compatibility
Claude Code 2.1.220+.
description
Agent orchestration patterns for agentic loops, multi-agent coordination, alternative frameworks, and multi-scenario workflows. Use when building autonomous agent loops, coordinating multiple agents, evaluating CrewAI/AutoGen/Swarm, or orchestrating complex multi-step scenarios.
Comprehensive patterns for building and coordinating AI agents -- from single-agent reasoning loops to multi-agent systems and framework selection. Coordination and multi-scenario categories have individual rule files in rules/ loaded on-demand; loop and framework tutorials live upstream (see Upstream coverage), with house defaults in references/ork-delta.md.
CC native /workflows (2.1.154): Claude Code now ships dynamic workflows — ask Claude to create a workflow and it orchestrates tens-to-hundreds of agents in the background; view runs with /workflows. This is complementary to the patterns here: use CC /workflows for large-scale, fire-and-forget background fan-out (you check back later); use the bounded foreground Agent Teams / Task-tool patterns below when ≤8 agents must coordinate within a single skill invocation via shared memory (handoff files, mesh messaging). Different scale, not a replacement.
Ask only when genuinely blocked (CC 2.1.154): CC now reserves the multiple-choice question prompt for decisions it genuinely cannot make itself, rather than asking when it already has enough context to proceed. When orchestrating agents, don't gate progress on an AskUserQuestion the lead can resolve from available context — reserve prompts for true branch points (irreversible actions, missing requirements). This complements ork's voice-friendly decision guidance.
Total: 4 rules across 4 categories. Loop and framework tutorials moved to first-party sources; the rescued house defaults live in .
references/ork-delta.md
Quick Start
# ReAct agent loopasyncdefreact_loop(question: str, tools: dict, max_steps: int = 10) -> str:
history = REACT_PROMPT.format(tools=list(tools.keys()), question=question)
for step inrange(max_steps):
response = await llm.chat([{"role": "user", "content": history}])
if"Final Answer:"in response.content:
return response.content.split("Final Answer:")[-1].strip()
if"Action:"in response.content:
action = parse_action(response.content)
result = await tools[action.name](*action.args)
history += f"\nObservation: {result}\n"return"Max steps reached without answer"
# Supervisor with fan-out/fan-inasyncdefmulti_agent_analysis(content: str) -> dict:
agents = [("security", security_agent), ("perf", perf_agent)]
tasks = [agent(content) for _, agent in agents]
results = await asyncio.gather(*tasks, return_exceptions=True)
returnawait synthesize_findings(results)
Agent Loops
Patterns for autonomous LLM reasoning: ReAct (Reasoning + Acting), Plan-and-Execute with replanning, self-correction loops, and sliding-window memory management.
Key decisions: Max steps 5-15, temperature 0.3-0.7, memory window 10-20 messages.
Multi-Agent Coordination
Fan-out/fan-in parallelism, supervisor routing with dependency ordering, conflict resolution (confidence-based or LLM arbitration), result synthesis, and CC Agent Teams (mesh topology for peer messaging in CC 2.1.33+).
Key decisions: 3-8 specialists, parallelize independent agents, use Task tool (star) for simple work, Agent Teams (mesh) for cross-cutting concerns.
Alternative Frameworks
CrewAI hierarchical crews with Flows (1.8+), OpenAI Agents SDK handoffs and guardrails (0.12+), Microsoft Agent Framework (AutoGen + SK merger), GPT-5.2-Codex for long-horizon coding, and AG2 for open-source flexibility.
Key decisions: Match framework to team expertise + use case. LangGraph for state machines, CrewAI for role-based teams, OpenAI SDK for handoff workflows, MS Agent for enterprise compliance.
Multi-Scenario
Orchestrate a single skill across 3 parallel scenarios (simple/medium/complex) with progressive difficulty scaling (1x/3x/8x), milestone synchronization, and cross-scenario result aggregation.
Key decisions: Free-running with checkpoints, always 3 scenarios, 1x/3x/8x exponential scaling, 30s/90s/300s time budgets.
Upstream coverage (do not restate)
Local tutorials for these topics were retired; consult the first-party source and keep only house deltas in references/ork-delta.md.
Multi-scenario state machine, architecture and skill-agnostic template deep-dives
Superseded in-skill by rules/scenario-orchestrator.md and rules/scenario-routing.md
References
references/ork-delta.md - House defaults and dated decisions rescued from retired tutorials
references/framework-comparison.md - Condensed framework decision matrix and use-case table
references/langgraph-implementation.md - LangGraph 1.2+ implementation of the multi-scenario orchestrator
references/claude-code-instance-management.md - Running 3 parallel Claude Code instances for scenario demos
Key Decisions
Decision
Recommendation
Single vs multi-agent
Single for focused tasks, multi for decomposable work
Max loop steps
5-15 (prevent infinite loops)
Agent count
3-8 specialists per workflow
Framework
Match to team expertise + use case
Topology
Task tool (star) for simple; Agent Teams (mesh) for complex
Scenario count
Always 3: simple, medium, complex
Common Mistakes
No step limit in agent loops (infinite loops)
No memory management (context overflow)
No error isolation in multi-agent (one failure crashes all)
Note (CC 2.1.161): parallel tool calls now fail independently — a failed Bash no longer cancels siblings in the same batch. This caveat still applies at the agent-orchestration level, not to tool batches; claude agents rows now show done/total for fanned-out work.
Note (CC 2.1.157): claude agents honors the agent field in settings.json for dispatched sessions; --agent <name> overrides it — pin the agent type explicitly when dispatching.
Missing synthesis step (raw agent outputs not useful)
Mixing frameworks in one project (complexity explosion)
Using Agent Teams for simple sequential work (use Task tool)
Sequential instead of parallel scenarios (defeats purpose)
