AI agent security testing
Conduct security testing on AI agent systems to detect protocol layer attacks such as tool poisoning and capability upgrades
0Uses25Likes31Views
What this skill does
Conduct security testing on AI agent systems to detect protocol layer attacks such as tool poisoning and capability upgrades
- Clarifies goals, inputs, and success criteria
- Preserves evidence and marks uncertain boundaries
- Produces structured, editable results