DeepSeek Harness Plugin Hub
← Plugins

dsh-skill-optimization

Skill Optimization

Treat skills as trainable parameters: train agent skills like neural networks (epochs/batchsize/learning rate/validation gates, without modifying model weights)—a rollout→reflect→aggregate→select→update→evaluate loop; accept candidate edits only when they strictly improve the held-out validation sco

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:satan9394/dsh-skill-optimization#eb41d9f3403ac48c0ebbbb5389ac11e8dd6ff2c0

Description

Treat skills as trainable parameters: train agent skills like neural networks (epochs/batchsize/learning rate/validation gates, without modifying model weights)—a rollout→reflect→aggregate→select→update→evaluate loop; accept candidate edits only when they strictly improve the held-out validation score; text learning-rate budget; zero inference-time model calls; deploy a compact best_skill.md. Inspired by microsoft/SkillOpt (MIT).

Compatibility and provenance

Skill Optimization is published as dsh-skill-optimization and currently resolves to version 0.1.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated

Versions

0.1.0stable