Skip to content

Conversation

@lorenss-m
Copy link
Contributor

@lorenss-m lorenss-m commented Jan 8, 2026

Note

Introduces hierarchical agent orchestration and gateway routing, plus MCP serving improvements and docs.

  • New AgentTool: Wrap scenarios as tools with eval-only param filtering and trace continuity; exported in hud.tools with tests
  • Gateway integration: create_agent() factory, gateway.py client builder, and model resolver.py; CLI eval routes by provider; comprehensive tests
  • Environment MCP serving: Auto-connect on run, override list/call to include connector tools; adds tests; docs add serve()/http_app() usage
  • Task/Eval semantics: Task.args now optional (template vs runnable); loaders default args to {}; eval context validates args and scenario setup uses {}; runners normalize AgentType usage; telemetry/tests adjusted
  • Schema tweak: Strip format in strict JSON schema
  • Docs: New "Ops Diagnostics Agent" cookbook; expands Tools (AgentTool) and Environment (serving) references; updates docs.json
  • Version: Bump to 0.5.3 (code and tests)

Written by Cursor Bugbot for commit cd0cc40. This will update automatically on new commits. Configure here.

- Add Scenario.as_agent_tool() for creating tools that spawn fresh agents
- Add Scenario.from_remote() for remote scenario handles via MCP
- Add hud.scenario() module-level helper
- Add create_agent() factory in hud/agents for programmatic agent creation
- Resolve stash merge conflicts (taskset_id naming)
@lorenss-m lorenss-m merged commit 227ea8a into main Jan 9, 2026
9 checks passed
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.

2 participants