LibraryAI and agent development
SkillOpt
Optimize reusable natural-language agent skills with trajectory-driven edits and validation gates.
- Stars
- 12,683
- Forks
- 1,178
- Updated
- Updated Jul 14, 2026
MIT
Optimize reusable natural-language agent skills with trajectory-driven edits and validation gates.
SkillOpt treats a skill document as the trainable state of an otherwise frozen agent. An optimizer model converts scored trajectories into bounded add, delete, or replace edits, and a candidate is accepted only when it strictly improves a held-out validation score. The resulting best_skill.md can be deployed without changing model weights or adding extra model calls at inference time, giving researchers and engineering teams a controlled way to improve reusable agent instructions.
Resource types
Use cases
Platforms
Add searchable long-term memory to AI assistants and agents.
Build lightweight agents that act through Python code or tool calls.
Runtime
Capabilities
Public GitHub facts last synced Jul 14, 2026.
Add Claude Code-inspired tools and skills to Spring AI applications.
Turns GitHub repositories into executable software-engineering tasks for training SWE agents.