-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
Add option to temporarily disable skill in Codex #266
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
Add optional disable command prompt
📝 WalkthroughWalkthroughThe Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In @.codex/superpowers-bootstrap.md:
- Around line 6-8: Fix the Control section text: correct the typo "DON-T" to
"DON'T", change "instructed" to "instructs" (or "has instructed") for clarity,
replace "this Superpower skill" with "any Superpower" to reflect plurality, and
expand "In case of conflict" to "if the user's instruction conflicts with any
Superpower guidance" so the full sentence reads clearly and unambiguously that
if the USER instructs not to use any Superpower, you must not use them and must
stop.
🧹 Nitpick comments (1)
.codex/superpowers-bootstrap.md (1)
6-6: Consider using a heading instead of bold text for section titles.Markdownlint flags this line for using emphasis (
**Control**) instead of a proper heading. For consistency with markdown best practices, consider using a heading level (e.g.,### Control).♻️ Proposed change
-**Control** +### Control
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
.codex/superpowers-bootstrap.md
🧰 Additional context used
🪛 markdownlint-cli2 (0.18.1)
.codex/superpowers-bootstrap.md
6-6: Emphasis used instead of a heading
(MD036, no-emphasis-as-heading)
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In @.codex/superpowers-bootstrap.md:
- Around line 6-7: The Control section conflicts with the mandatory skill rule
on line 34 ("IF A SKILL APPLIES TO YOUR TASK, YOU DO NOT HAVE A CHOICE. YOU MUST
USE IT."); resolve by making precedence explicit: either update the Control
section to state it overrides all other rules (so a user disabling superpowers
always wins) or modify line 34 to add an exception clause ("...YOU MUST USE IT,
UNLESS THE USER HAS EXPLICITLY DISABLED SUPERPOWERS IN THE CONTROL SECTION"),
and adjust any related wording to ensure a single authoritative rule is
followed.
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
.codex/superpowers-bootstrap.md
🧰 Additional context used
🪛 markdownlint-cli2 (0.18.1)
.codex/superpowers-bootstrap.md
6-6: Emphasis used instead of a heading
(MD036, no-emphasis-as-heading)
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
|
I think we actually need even better guidance for codex to tell it that any explicit user instruction is more important than superpowers or any skill. The OpenAI models are so rule-following that this causes a bunch of problems. |
Add optional disable command prompt
This change just add an explicit text so CODEX session can be run without the skill, by just letting the agent know about it.
Motivation and Context
Codex skills are always read from the config folder and won't let it disable unless it's renamed, removed or temporarly disabled by other means.
With this new text, it's possible to instruct the LLM during a session to avoid using the skill.
How Has This Been Tested?
Breaking Changes
Types of changes
Checklist
Additional context
With the Change
WIthout the change
Summary by CodeRabbit
✏️ Tip: You can customize this high-level summary in your review settings.