Implement recursive agent delegation through delegate_task tool calls
Run delegated sub-agents, feed their results back as tool_result, and handle unknown agents, failures, and circular delegation.
claude-code-by-agents-recursive-delegation
Normalized profile
What this task asks for
Run delegated sub-agents, feed their results back as tool_result, and handle unknown agents, failures, and circular delegation. Trigger delegation through `delegate_task` with an `agent_id` and delegated instructions. Run the selected sub-agent and feed one JSON `tool_result` containing its accumulated text back to the delegating agent. Match the streamed tool-use ID to `tool_result.tool_use_id`. Report unknown agents with both a stream error and an error tool result that names the requested agent. Report sub-agent failures only through an error tool result. Report circular delegation with a stream-level error whose message mentions `circular`.
Intent
- feature implementation
Work surfaces
- network service
Technology
- Claude Code by Agents
- JSON
Expected artifacts
- repository patch
What makes it difficult
- recursive delegation and cross-agent error propagation
Published evidence
Model × harness outcomes
Counts are tied to these exact configurations. Harnesses are not treated as equivalent across benchmarks.
- Models
- 13
- Configurations
- 41
- Trials
- 163
Showing 41 of 41 configurations
| Model | Harness | Effort | Result | Passes |
|---|---|---|---|---|
| claude-fable-5vertex_ai | mini-swe-agentpublished configuration | low | Mixed | 2/4 |
| claude-fable-5vertex_ai | mini-swe-agentpublished configuration | medium | Mixed | 1/4 |
| claude-fable-5vertex_ai | mini-swe-agentpublished configuration | high | Mixed | 1/4 |
| claude-fable-5vertex_ai | mini-swe-agentpublished configuration | xhigh | Mixed | 1/4 |
| claude-fable-5vertex_ai | mini-swe-agentpublished configuration | max | No passes | 0/4 |
| claude-opus-4-8anthropic | mini-swe-agentpublished configuration | low | Mixed | 3/4 |
| claude-opus-4-8anthropic | mini-swe-agentpublished configuration | medium | Mixed | 3/4 |
| claude-opus-4-8anthropic | mini-swe-agentpublished configuration | high | Mixed | 1/4 |
| claude-opus-4-8anthropic | mini-swe-agentpublished configuration | xhigh | Mixed | 1/4 |
| claude-opus-4-8anthropic | mini-swe-agentpublished configuration | max | Mixed | 1/4 |
| claude-sonnet-4-6anthropic | mini-swe-agentpublished configuration | high | Mixed | 3/4 |
| claude-sonnet-5anthropic | mini-swe-agentpublished configuration | low | Mixed | 3/4 |
| claude-sonnet-5anthropic | mini-swe-agentpublished configuration | medium | No passes | 0/4 |
| claude-sonnet-5anthropic | mini-swe-agentpublished configuration | high | Mixed | 1/4 |
| claude-sonnet-5anthropic | mini-swe-agentpublished configuration | xhigh | Mixed | 1/4 |
| claude-sonnet-5anthropic | mini-swe-agentpublished configuration | max | No passes | 0/4 |
| gemini-3-1-pro-previewvertex_ai | mini-swe-agentpublished configuration | high | Mixed | 2/4 |
| gemini-3-5-flashvertex_ai | mini-swe-agentpublished configuration | medium | Mixed | 3/4 |
| glm-5-2zai | mini-swe-agentpublished configuration | high | Mixed | 2/4 |
| glm-5-2zai | mini-swe-agentpublished configuration | max | Mixed | 2/4 |
| gpt-5-4openai | mini-swe-agentpublished configuration | xhigh | All passed | 4/4 |
| gpt-5-5openai | mini-swe-agentpublished configuration | low | Mixed | 1/4 |
| gpt-5-5openai | mini-swe-agentpublished configuration | medium | Mixed | 1/4 |
| gpt-5-5openai | mini-swe-agentpublished configuration | high | Mixed | 1/4 |
| gpt-5-5openai | mini-swe-agentpublished configuration | xhigh | Mixed | 2/4 |
| gpt-5-6-lunaopenai | mini-swe-agentpublished configuration | low | Mixed | 2/4 |
| gpt-5-6-lunaopenai | mini-swe-agentpublished configuration | medium | Mixed | 1/4 |
| gpt-5-6-lunaopenai | mini-swe-agentpublished configuration | high | Mixed | 2/4 |
| gpt-5-6-lunaopenai | mini-swe-agentpublished configuration | xhigh | Mixed | 2/4 |
| gpt-5-6-lunaopenai | mini-swe-agentpublished configuration | max | All passed | 4/4 |
| gpt-5-6-solopenai | mini-swe-agentpublished configuration | low | All passed | 4/4 |
| gpt-5-6-solopenai | mini-swe-agentpublished configuration | medium | All passed | 4/4 |
| gpt-5-6-solopenai | mini-swe-agentpublished configuration | high | All passed | 4/4 |
| gpt-5-6-solopenai | mini-swe-agentpublished configuration | xhigh | Mixed | 3/4 |
| gpt-5-6-solopenai | mini-swe-agentpublished configuration | max | Mixed | 2/4 |
| gpt-5-6-terraopenai | mini-swe-agentpublished configuration | low | All passed | 3/3 |
| gpt-5-6-terraopenai | mini-swe-agentpublished configuration | medium | All passed | 4/4 |
| gpt-5-6-terraopenai | mini-swe-agentpublished configuration | high | All passed | 4/4 |
| gpt-5-6-terraopenai | mini-swe-agentpublished configuration | xhigh | All passed | 4/4 |
| gpt-5-6-terraopenai | mini-swe-agentpublished configuration | max | All passed | 4/4 |
| kimi-k2-7-codeopenrouter | mini-swe-agentpublished configuration | default | Mixed | 2/4 |