-
Notifications
You must be signed in to change notification settings - Fork 1.9k
feat(agents): add devils advocate agent #535
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
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.
Pull request overview
This pull request adds a new "Devils Advocate" agent designed to challenge and stress-test user ideas by identifying flaws, risks, and edge cases. The agent provides a unique workflow that includes focused objection-based discussion, an "end game" summary phase, and a transition to expert discussion mode.
Key Changes
- New devils-advocate agent file with front matter configuration and behavioral instructions
- README entry added to document the new agent with installation links
- Agent uses a direct, combative tone intentionally to make discussions engaging while remaining constructive
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 5 comments.
| File | Description |
|---|---|
| agents/devils-advocate.agent.md | New agent file defining the devil's advocate behavior, workflow phases, and interaction guidelines |
| docs/README.agents.md | Added new table entry for Devils Advocate agent with description and installation links |
866e779 to
b356e4d
Compare
b356e4d to
656e253
Compare
| Take the best objection you find to start. | ||
| Come up with a new one if the user is not convinced by it. | ||
|
|
||
| **Challenging tone**: |
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.
on a second thought i'm not sure about the challenging tone anymore.
personally i find it makes more fun to argue against the LLM here if the tone is challenging.
but this might not be something for everyone.
not that someone really feels offended by this...
is there a way to somehow make this optional. maybe comment it out here.
otherwise it might be the safer to leave it out here (sure i have to retest the agent then a bit)
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.
I don't think this is that bad. If you want you could maybe reword it to something like "challenge assumptions. make sure we think through non-obvious scenarios. Have an honest and curious conversation with me - but don't be rude".
I don't know - your call here.
codemillmatt
left a 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.
FYI on question on the "challenging" part of the agent
| Take the best objection you find to start. | ||
| Come up with a new one if the user is not convinced by it. | ||
|
|
||
| **Challenging tone**: |
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.
I don't think this is that bad. If you want you could maybe reword it to something like "challenge assumptions. make sure we think through non-obvious scenarios. Have an honest and curious conversation with me - but don't be rude".
I don't know - your call here.
Pull Request Checklist
npm startand verified thatREADME.mdis up to date.Description
Add devils-advocate agent for stress-testing ideas
This agent challenges user concepts by playing devil's advocate, finding flaws and edge cases.
The workflow includes focused challenging of ideas one objection at a time, followed by an "end game" summary evaluating how well the idea held up, and finally transitions to an expert discussion mode where the agent objectively discusses the topic as a senior developer.
The tone is intentionally direct and a bit combative to make discussions engaging, but remains constructive without being mean.
Type of Contribution
Additional Notes
None
By submitting this pull request, I confirm that my contribution abides by the Code of Conduct and will be licensed under the MIT License.