feat: allow model invocation for adversarial-review command (#156)#157
feat: allow model invocation for adversarial-review command (#156)#157D2758695161 wants to merge 2 commits intoopenai:mainfrom
Conversation
![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: 5c18d340a6
ℹ️ About Codex in GitHub
Codex has been enabled to automatically 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 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| model: request.model, | ||
| effort: request.effort, | ||
| sandbox: request.write ? "workspace-write" : "read-only", | ||
| sandbox: request.fullAccess ? "danger-full-access" : (request.write ? "workspace-write" : "read-only"), |
There was a problem hiding this comment.
handleTask now accepts --full-access, and executeTaskRun switches to danger-full-access when request.fullAccess is truthy, but no call site ever sets that property (foreground call to executeTaskRun and background buildTaskRequest only pass write). As a result, --full-access is silently ignored in both modes, so tasks that require unrestricted sandboxing still run as workspace-write/read-only.
Useful? React with 👍 / 👎.
) Remove `disable-model-invocation: true` from adversarial-review, cancel, result, and status commands so Claude can route explicit user requests and other skills/workflows can invoke them. The flag is intentionally kept on `review` to prevent proactive Codex spend on the heavyweight native review. An explicit guardrail is added to adversarial-review telling Claude to only run it when explicitly asked. Also fixes a pre-existing test regression from openai#168 where the quoted `$ARGUMENTS` pattern was not reflected in the test regex. Closes openai#211 Supersedes openai#156 openai#157 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1 participant
Summary
Removes
disable-model-invocation: truefrom theadversarial-reviewcommand frontmatter.Problem
The
disable-model-invocationflag prevents skills from invoking/codex:adversarial-reviewvia the Skill tool, blocking workflow composition.Solution
Remove
disable-model-invocation: truefrom the frontmatter. The command has two independent safety layers:allowed-toolsrestricts to Read/Glob/Grep/Bash/AskUserQuestion - no edit accessChanges
plugins/codex/commands/adversarial-review.md: Removeddisable-model-invocation: truefrom frontmatterCloses #156