fix(ce-work-beta): defer model and reasoning effort to Codex config#704
Merged
fix(ce-work-beta): defer model and reasoning effort to Codex config#704
Conversation
Drop the baked-in `gpt-5.4` model default and `high` reasoning-effort default from `work_delegate_model` and `work_delegate_effort`. The skill no longer passes `-m` or `-c 'model_reasoning_effort=...'` to `codex exec` when the user has not set these in `.compound-engineering/config.local.yaml` -- Codex then resolves from `~/.codex/config.toml` (and ultimately the CLI's own default). Avoids two ongoing problems: chasing new model names in skill content as Codex evolves, and silently overriding whatever default the user configured globally for Codex. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
![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: 649fc6a475
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".
The previous edit dropped hard defaults for `work_delegate_model` and `work_delegate_effort` but left the global "fall through to the hard default" rule unmodified. A typo such as `work_delegate_effort: hgh` had no deterministic fallback and could lead the orchestrator to pass an invalid effort flag through to `codex exec`. State explicitly that unrecognized or unparseable values for these two keys resolve to unset, and that the corresponding flag must be omitted from the `codex exec` invocation rather than substituted with the invalid value. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Merged
michaelvolz
pushed a commit
to michaelvolz/compound-engineering-plugin-windows-version
that referenced
this pull request
Apr 28, 2026
…veryInc#704) Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
work_delegate_modelandwork_delegate_effortare now optional with no baked-in default. When unset, the skill omits-mand-c 'model_reasoning_effort=...'fromcodex execso Codex resolves from~/.codex/config.toml(and ultimately the CLI's own default)..compound-engineering/config.local.yamlif they want to deviate from their global Codex config — that path is unchanged.gpt-5.4/highas our defaults.The previous behavior forced two ongoing problems: chasing new model names inside skill content as Codex evolves, and silently overriding whatever default the user already configured for Codex globally.
Files changed
plugins/compound-engineering/skills/ce-work-beta/SKILL.md— config keys reframed as optional; resolved state may be unset.plugins/compound-engineering/skills/ce-work-beta/references/codex-delegation-workflow.md—-mand-clines lifted out of thecodex exectemplate into a "Conditional flags" block that inserts each line only when the corresponding skill-state value is set..compound-engineering/config.local.example.yaml,plugins/compound-engineering/skills/ce-setup/references/config-template.yaml— comments updated.Stable/beta sync: not propagating to
ce-work— codex delegation only exists in beta.Test plan
work_delegate_model/work_delegate_effortin config: confirmcodex execinvocation has neither-mnor-c 'model_reasoning_effort=...'(Codex picks up~/.codex/config.toml).work_delegate_model: gpt-5.4set: confirm-m "gpt-5.4"is included.work_delegate_effort: mediumset: confirm-c 'model_reasoning_effort="medium"'is included.bun run release:validatepasses (already verified locally).🤖 Generated with Claude Code