Skip to content

feat(agent): record hierarchical handoff evidence#317

Merged
JaimeCernuda merged 1 commit into
developfrom
feat/hierarchical-benchmark-evidence-20260524
May 24, 2026
Merged

feat(agent): record hierarchical handoff evidence#317
JaimeCernuda merged 1 commit into
developfrom
feat/hierarchical-benchmark-evidence-20260524

Conversation

@JaimeCernuda
Copy link
Copy Markdown
Collaborator

Summary

  • add first-class ExpertHandoff provenance to CLIO run traces
  • expose expert_handoffs on GACT assistant message metadata and stress audit logs
  • remove a remaining closed data|analysis|visualization expert check so nested experts surface tool errors and ARC invocations like other agents
  • update hierarchical benchmark docs/TASK notes to require handoff evidence, not just final route labels

Verification

  • uv run ruff check src/clio_agent/harness.py src/clio_agent/agent.py src/clio_agent/gact/app.py tests/test_core/test_agent_dispatch.py tests/test_gact/test_tools_called.py tests/test_stress_benchmark/test_local_scientific_workflows.py
  • uv run pytest tests/test_core/test_agent_dispatch.py tests/test_core/test_agent_planner.py::TestBuildCapabilitiesContext tests/test_core/test_agent_planner.py::TestRunAgentLoop::test_ndp_tool_action_is_promoted_to_nested_catalog_expert tests/test_gact/test_agents_catalog.py tests/test_gact/test_tools_called.py tests/test_tools/test_ndp_server.py tests/test_tools/test_sac_server.py tests/test_experts/test_analysis_expert.py tests/test_experts/test_data_expert.py -q -> 110 passed
  • uv run pytest tests/ -q -> 1154 passed, 37 skipped

@JaimeCernuda JaimeCernuda merged commit 8ac0375 into develop May 24, 2026
1 check failed
@JaimeCernuda JaimeCernuda deleted the feat/hierarchical-benchmark-evidence-20260524 branch May 24, 2026 04:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant