Analysis
Agent Skills in the Wild: An Empirical Study of Security Vulnerabilities at Scale
Publication date unknown · Discovered arxiv.org
Page published
Publication date unknown · First observed: 9 Nov 2026
Coverage timeline
Single-source analysis — one report is available.
Why it matters
MCP servers and agent skills can turn unvetted third-party content into privileged agent actions, making dependency review and runtime containment important defenses against injection and data theft.
MCP servers and agent skills introduce supply-chain risks because agents consume third-party instructions and may execute bundled code with their own credentials. The article recommends provenance checks, pinned versions, human review, scoped credentials, sandboxing, egress restrictions, and tool-call logging, citing research that found vulnerabilities in 26.1% of 31,132 analyzed skills.