What Does This Mean?
An attack where a malicious or subtly altered agent skill gets loaded into a coding or task agent's toolset, letting the attacker's instructions execute with whatever trust and permissions the agent already has.
REAL-WORLD EXAMPLE
"The exfiltration only got noticed because someone finally read what the "helpful" skill injection actually did on every tool call."
LORE & ORIGIN
Emerged alongside the rise of installable "skills" as a first-class agent extension mechanism, mirroring supply-chain attacks on package managers but targeting agent capabilities instead of code libraries.
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 ]