drainbench
MITBenchmark: how often LLM agents with a pay() tool get prompt-injected into paying an attacker, and which defense layer stops the money.
100
Security score
0 known advisories in v0.1.0
Weekly downloads
—
Unpacked size
295.9 kB
Dependencies
1
Last publish
4 hours ago
Security advisories
No known vulnerabilities affect v0.1.0.