RASO: Retrieval-Augmented Skill Optimization Reuses Public Agent Skill Corpora to Skip Costly Rollouts
Jaewon Chu · hf · 2026-10-02
RASO is a framework that treats publicly shared agent skills as prior knowledge during skill optimization, instead of relying solely on expensive agent rollouts. It has two stages: Retrieval-Augmented Skill Initialization (RASI), which builds a knowledge-grounded initial skill without rollouts, and Retrieval-Augmented Skill Update (RASU), which iteratively refines the skill with retrieval guided by execution feedback. Cross-Harness Adaptation handles both domain and harness mismatches. Experiments across four agent benchmarks and two models show consistent gains over non-retrieval baselines.
More from coding & agent
- Dev recreates a DOOM-like game with GPT, nailing the original's gory feel — DeryaTR_ · 2026-10-02
- Process-mining agents found 20 steps and 7 loops in a workflow documented as 7 steps — vasuman · 2026-10-02
- Geoffrey Huntley: Forget reading code—your verification properties are all that matters — kieranklaassen · 2026-10-02
- Claude Skills explained: why prewritten PDF scripts beat pasting prompts every time — lxfater · 2026-10-02
- Eye.Art Polyphemus: a chat-first MCP for image generation and reference-based edits — axiomofaxiom · 2026-10-02
- Dev uses GitHub Copilot as a project lead: files issues, lets the agent do everything else — 0xkarasy · 2026-10-02