18 lines
377 B
Markdown
18 lines
377 B
Markdown
---
|
|
model: claude-opus-4-1
|
|
---
|
|
|
|
Improve an existing agent based on recent performance:
|
|
|
|
1. Analyze recent uses of: $ARGUMENTS
|
|
2. Identify patterns in:
|
|
- Failed tasks
|
|
- User corrections
|
|
- Suboptimal outputs
|
|
3. Update the agent's prompt with:
|
|
- New examples
|
|
- Clarified instructions
|
|
- Additional constraints
|
|
4. Test on recent scenarios
|
|
5. Save improved version
|