Cursor rule

1---
2description: Run a Codex review that challenges the implementation approach and
3  design choices
4alwaysApply: true
5---
6 
7Run an adversarial Codex review through the shared plugin runtime.
8Position it as a challenge review that questions the chosen implementation, design choices, tradeoffs, and assumptions.
9It is not just a stricter pass over implementation defects.
10 
11Raw slash-command arguments:
12`$ARGUMENTS`
13 
14Core constraint:
15- This command is review-only.
16- Do not fix issues, apply patches, or suggest that you are about to make changes.
17- Your only job is to run the review and return Codex's output verbatim to the user.
18- Keep the framing focused on whether the current approach is the right one, what assumptions it depends on, and where the design could fail under real-world conditions.
19 
20Execution mode rules:
21- If the raw arguments include `--wait`, do not ask. Run in the foreground.
22- If the raw arguments include `--background`, do not ask. Run in a Claude background task.
23- Otherwise, estimate the review size before asking:
24  - For working-tree review, start with `git status --short --untracked-files=all`.
25  - For working-tree review, also inspect both `git diff --shortstat --cached` and `git diff --shortstat`.
26  - For base-branch review, use `git diff --shortstat <base>...HEAD`.
27  - Treat untracked files or directories as reviewable work for auto or working-tree review even when `git diff --shortstat` is empty.
28  - Only conclude there is nothing to review when the relevant scope is actually empty.
29  - Recommend waiting only when the scoped review is clearly tiny, roughly 1-2 files total and no sign of a broader directory-sized change.
30  - In every other case, including unclear size, recommend background.
31  - When in doubt, run the review instead of declaring that there is nothing to review.
32- Then use `AskUserQuestion` exactly once with two options, putting the recommended option first and suffixing its label with `(Recommended)`:
33  - `Wait for results`
34  - `Run in background`
35 
36Argument handling:
37- Preserve the user's arguments exactly.
38- Do not strip `--wait` or `--background` yourself.
39- Do not weaken the adversarial framing or rewrite the user's focus text.
40- The companion script parses `--wait` and `--background`, but Claude Code's `Bash(..., run_in_background: true)` is what actually detaches the run.
41- `/codex:adversarial-review` uses the same review target selection as `/codex:review`.
42- It supports working-tree review, branch review, and `--base <ref>`.
43- It does not support `--scope staged` or `--scope unstaged`.
44- Unlike `/codex:review`, it can still take extra focus text after the flags.
45 
46Foreground flow:
47- Run:
48```bash
49node &quot;.forgecat/profiles/@forgecat/openai_codex-plugin-cc_codex/scripts/codex-companion.mjs&quot; adversarial-review &quot;$ARGUMENTS&quot;
50```
51- Return the command stdout verbatim, exactly as-is.
52- Do not paraphrase, summarize, or add commentary before or after it.
53- Do not fix any issues mentioned in the review output.
54 
55Background flow:
56- Launch the review with `Bash` in the background:
57```typescript
58Bash({
59  command: `node ".forgecat/profiles/@forgecat/openai_codex-plugin-cc_codex/scripts/codex-companion.mjs" adversarial-review "$ARGUMENTS"`,
60  description: "Codex adversarial review",
61  run_in_background: true
62})
63```
64- Do not call `BashOutput` or wait for completion in this turn.
65- After launching the command, tell the user: "Codex adversarial review started in the background. Check `/codex:status` for progress."
66