# Agent Skill Scanners Disagree So Sharply That No Single Verdict Is Usable

Across three scanners applied to a public agent skill registry, any pair agreed on at most 10.4% of their combined detections, and 81.9% of flagged skills were caught by one scanner alone.

- Published: 2026-09-14T06:22:55.303Z
- Canonical: https://polylog.news/ai/2026-09-14/agent-skill-scanners-disagree-so-sharply-that-no-single-verd
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [arXiv](https://arxiv.org/abs/2609.12001)

A preprint posted on September 14, Scan the Skill, Govern the Action, starts from a measurement that undermines how agent skill registries currently work. Security scanning of published skills, the packaged instructions and tools that agent…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-09-14/agent-skill-scanners-disagree-so-sharply-that-no-single-verd (subscription information: https://polylog.news/pricing).