Update Codex sub-agent model assignments#72
Conversation
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
WalkthroughChangesUpdate Codex model selections for four subagents, raise Wyvern and Scribe reasoning effort to high, synchronize installer output, and revise tests to assert the new configuration contracts. Codex configuration refresh
Suggested reviewers: Poem
🚥 Pre-merge checks | ✅ 20✅ Passed checks (20 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
|
@coderabbitai review |
✅ Action performedReview finished.
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
tests/test_subagent_definitions.py (1)
41-68: 📐 Maintainability & Code Quality | 🟠 Major | ⚡ Quick winParameterise the duplicate Codex contract tests. Collapse the Wyvern and Scribe checks into one
pytest.mark.parametrizecase; only the subagent name and sandbox mode differ.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@tests/test_subagent_definitions.py` around lines 41 - 68, Replace the duplicate tests test_wyvern_codex_subagent_uses_luna_model and test_scribe_codex_subagent_uses_luna_model with one pytest.mark.parametrize-driven test covering each subagent name and its expected sandbox mode. Keep the shared model and reasoning_effort assertions, and parameterize only the differing subagent name and sandbox_mode values.Source: Path instructions
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@tests/test_subagent_definitions.py`:
- Around line 41-68: Replace the duplicate tests
test_wyvern_codex_subagent_uses_luna_model and
test_scribe_codex_subagent_uses_luna_model with one
pytest.mark.parametrize-driven test covering each subagent name and its expected
sandbox mode. Keep the shared model and reasoning_effort assertions, and
parameterize only the differing subagent name and sandbox_mode values.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: a099b4a4-ee23-4c7a-8fb8-4f7d7e9a83a4
📒 Files selected for processing (3)
agents/subagents.ymlinstall-sub-agentstests/test_subagent_definitions.py
|
@coderabbitai review |
✅ Action performedReview finished.
|
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@tests/test_subagent_definitions.py`:
- Around line 41-64: Replace the duplicate test functions
test_wyvern_codex_subagent_uses_luna_model and
test_scribe_codex_subagent_uses_luna_model with one `@pytest.mark.parametrize`
test, supplying each agent name and expected sandbox mode as parameters. Reuse
the shared load_provider lookup and Luna model, reasoning_effort, and
sandbox_mode assertions while preserving each agent’s existing expectations.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: eb72a6b2-7ba4-45ed-bc29-4308107cf577
📒 Files selected for processing (3)
agents/subagents.ymlinstall-sub-agentstests/test_subagent_definitions.py
Move Wyvern, Scribe, and Scrutineer to `gpt-5.6-luna` with their role-specific reasoning effort. Move Alchemist to `gpt-5.6-terra` with medium effort, and pin all four contracts in regression tests.
Consolidate the duplicate Wyvern and Scribe checks while preserving each agent's sandbox-mode contract.
Parse the Wyvern and Scribe TOML emitted by `install-sub-agents` and pin their Luna model and high reasoning-effort settings.
812664f to
eb667a2
Compare
|
@coderabbitai review |
✅ Action performedReview finished.
|
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@tests/test_subagent_definitions.py`:
- Around line 41-44: Update the values container in the pytest.mark.parametrize
decorator to use a list of tuples instead of a tuple of tuples, while preserving
the existing test cases and parameter names.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: 6976f63a-d609-4d68-aedb-af0f87dec54c
📒 Files selected for processing (4)
agents/subagents.ymlinstall-sub-agentstests/test_rust_entrypoints.pytests/test_subagent_definitions.py
Match the pytest parametrization style required by Ruff PT007 while preserving the existing agent cases and sandbox expectations.
2 participants
Summary
This branch updates the Codex-specific model assignments for the managed
sub-agents so that each role uses the requested GPT-5.6 model and reasoning
effort. It keeps the declarative manifest, direct installer output, and
regression contracts aligned.
Wyvern and Scribe now use
gpt-5.6-lunawith high reasoning effort,Scrutineer uses
gpt-5.6-lunawith medium effort, and Alchemist usesgpt-5.6-terrawith medium effort. Claude Code and Goose configurations areunchanged.
Review walkthrough
Validation
make all: passed; 187 tests and three snapshots passed, together with formatting, lint, typecheck, and spelling gates.git diff --check origin/main...HEAD: passed.References