Study finds Agent skill injection may lower Pass@2 rates
rohanpaul_ai · x · 2026-08-27
Research indicates that attaching a skill file to every prompt in a coding session is often a net loss. The WebDev-Skills-Bench shows that while skill injection raises token costs by at least 72%, it lowers the mean Pass@2 by 1.3 to 4.2 points across four models. Losses concentrate on easy early tasks, likely because skills lock in structural choices, turning recoverable mistakes into chain-terminating failures.
More from coding & agent
- Musk amplifies dev's verdict: Grok Bot easier to use than Codex and Claude — elonmusk · 2026-10-15
- Blob0 launches: a macOS file-manager-style desktop client for S3 and Cloudflare R2 — yihui_indie · 2026-10-15
- The 6-Layer AI Engineer Skill Stack for 2026: Vectors to Multi-Agent Ops — mdancho84 · 2026-10-15
- Building a Mechanical Engineering Agent Workflow in ChatGPT: 30-Node Canvas Shows Promise — burhop · 2026-10-15
- Open-source Harness Engineering course hits 19k stars, shows $9 vs $200 gap between failing and working agent setups — thisguyknowsai · 2026-10-15
- A 6-person team shipped healthcare prior-auth AI agents in 3 months — here's the playbook — alex_verem · 2026-10-15