Planted 3 backdoors in my code; the agent swarm found 27, half noise
Source
github.com
Author
zkvimuon
Date
Why it matters
Shows how far an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → swarm's security findings can be trusted: the audit was re-checked against code, surfacing one serious sandboxAn isolated environment where AI-generated code or agent actions run without being able to touch anything real.Full definition → escape and many lesser issues, and separating real findings from noise.
Terms in this piece · Glossary
sandbox — An isolated environment where AI-generated code or agent actions run without being able to touch anything real.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
