v3.4.0 #200
Unanswered
mariocandela
asked this question in
Q&A
v3.4.0
#200
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
What's Changed
Add and configure decoy MCP tools to detect prompt injection attacks against LLM agents. Collect attack prompts in real time and use them for guardrail fine-tuning.
Full Changelog: v3.3.14...v3.4.0
This discussion was created from the release v3.4.0.
Beta Was this translation helpful? Give feedback.
All reactions