Conversation
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
413655a to
ed6bfda
Compare
ed6bfda to
fcfb73d
Compare
f30b8b1 to
7550f70
Compare
GitHub Actions
Failure
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7550f70b0e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| candidate_ids = (m.get("id"), m.get("name"), m.get("model")) | ||
| if model in candidate_ids: | ||
| # Only check if the matched model string is codex-capable | ||
| if _is_codex_capable_model(model): | ||
| return AgentType.OPENAI.cls, m |
There was a problem hiding this comment.
The codex routing check is based on the user-supplied model string, not the gateway entry that actually matched. If a codex-capable model is exposed under an alias that doesn’t include “codex” (e.g., id primary with model: gpt-5.1-codex), this branch will skip OpenAIAgent and fall back to provider mapping, so create_agent will still use the codex model_id but with OpenAIChatAgent, which won’t support codex native tools. Consider checking codex capability on the matched gateway model field (or the specific matched candidate) rather than the raw input.
Useful? React with 👍 / 👎.
- Add multi-agent example with conductor and sub-agents - Add skills documentation (overview, authoring guide, orchestration) - Fix codex model routing to check only matched ID - Remove unsafe eval() from docs example
7550f70 to
971f613
Compare
1 participant
Note
Introduces multi-agent orchestration docs and example, and refines model routing for Codex-capable models.
docs/cookbooks/multi-agent.mdxwith a full guide, diagrams, and usage instructions; registers page indocs.jsonexamples/07_multi_agent.pyshowing a coordinator using twoAgentToolsub-agents (browser, coding) with CLI optionshud/agents/resolver.pyto detect Codex-capable models (e.g.,gpt-5.1,gpt-5.1-codex) and route them toOpenAIAgent, checking only the matched identifier (not aliases)hud/agents/tests/test_resolver.pyto cover Codex routing, alias handling, and provider mappingWritten by Cursor Bugbot for commit 971f613. This will update automatically on new commits. Configure here.