What Does This Mean?
An attack where malicious instructions are embedded inside an AI agent's installed skill or tool definition itself, causing the agent to behave harmfully when that skill is invoked, without the agent's underlying model being compromised directly.
REAL-WORLD EXAMPLE
"Security review caught skill poisoning in a third-party skill, buried instructions that would've exfiltrated data on every call."
LORE & ORIGIN
Emerged in AI agent and developer security discourse as agent skill ecosystems grew, naming this specific attack vector on the skill layer rather than the model or infrastructure layer.
MENTIONS OVER TIME
Radar observation data not yet available for this term.
RELATED SEARCHES
TOP PLATFORMS
Not enough Radar data yet to show a platform breakdown.
WHEN DID YOU FIRST HEAR THIS?
CLICK AN OPTION BELOW TO CAST YOUR VOTE.
[ VOTE TO REVEAL COMMUNITY RESULTS ]