lbeurerkellner

7 ms·

31.
by lbeurerkellner 2y ago
Thanks for crediting us :)
32.
by lbeurerkellner 2y ago
This would work in an ideal setting, however, in my experience it is not compatible with the general expectations we have for agentic systems. For instance, what about a simple user query like "Can you install this library?". In t
33.
by lbeurerkellner 2y ago
The post highlights and cites a few attack scenarios we originally described in a security note (tool poisoning, shadowing, MCP rug pull), published a few days ago [1]. I am the author of said blog post at Invariant Labs. Different from wha
34.
MCP Tool Poisoning: Taking over Your Favorite MCP Client(lbeurerkellner.github.io)
2 points by lbeurerkellner 2y ago|0 comments
35.
MCP Tool Poisoning: Taking over Your Favorite MCP Client(lbeurerkellner.github.io)
2 points by lbeurerkellner 2y ago|0 comments
36.
by lbeurerkellner 2y ago
Please be aware that MCP has severe security risks associated: https://invariantlabs.ai/blog/mcp-security-notification-tool... .
37.
MCP is all fun, until you add this one malicious MCP server and forget about it(twitter.com)
1 points by lbeurerkellner 2y ago|0 comments
38.
MCP Tool Poisoning: Taking over Your Favorite MCP Client(lbeurerkellner.github.io)
1 points by lbeurerkellner 2y ago|0 comments
39.
JSONSchemaBench: Generating Structured Outputs from Language Models(github.com)
1 points by lbeurerkellner 2y ago|0 comments
40.
by lbeurerkellner 2y ago
The diagrams are really cool. Congrats on the launch.
41.
Enhancing Browser Agent Safety with Guardrails(invariantlabs.ai)
1 points by lbeurerkellner 2y ago|0 comments
42.
Invariant: A security and bug scanner for agent traces(github.com)
1 points by lbeurerkellner 2y ago|0 comments
43.
Enhancing Browser Agent Safety with Guardrails(invariantlabs.ai)
1 points by lbeurerkellner 2y ago|0 comments
44.
by lbeurerkellner 2y ago
The security implications of this are very unclear it seems. Even the supervisor model can be fooled, and what if the agent just makes an honest mistake. It will be very interesting to see whether people are willing to let this actually go
45.
Security Scanner for AI Agent Traces: Invariant Analyzer(github.com)
1 points by lbeurerkellner 2y ago|0 comments
46.
playwright-computer-use: Let Claude control a web browser on your machine(github.com)
3 points by lbeurerkellner 2y ago|0 comments
47.
Invariant Agent Stack: A framework-less approach to robust agent development(github.com)
1 points by lbeurerkellner 2y ago|0 comments
48.
Show HN: Let Claude control a web browser on your machine(github.com)
3 points by lbeurerkellner 2y ago|0 comments
49.
Invariant Analyzer: Security scanner for AI agent trajectories(github.com)
6 points by lbeurerkellner 2y ago|0 comments
50.
Invariant Explorer: A tool for visualizing and exploring agent traces(github.com)
1 points by lbeurerkellner 2y ago|0 comments
51.
by lbeurerkellner 2y ago
Check out https://explorer.invariantlabs.ai/benchmarks/ for an interesting list of use cases for agents.
52.
Show HN: Try test-driven agent development in this holiday prompting challenge(invariantlabs.ai)
3 points by lbeurerkellner 2y ago|0 comments
53.
by lbeurerkellner 2y ago
Let us know if you can think of any benchmark, that you'd like to see added.
54.
Show HN: A registry of agent benchmarks (including many OSS agent trajectories)(explorer.invariantlabs.ai)
6 points by lbeurerkellner 2y ago|1 comments
55.
Explorer: A tool for visualizing and exploring agent traces(github.com)
1 points by lbeurerkellner 2y ago|0 comments
56.
Releasing Explorer and Testing: Visualize and Understand AI Agents(invariantlabs.ai)
1 points by lbeurerkellner 2y ago|0 comments
57.
Testing: Build better AI agents through debuggable unit testing(github.com)
1 points by lbeurerkellner 2y ago|0 comments
58.
by lbeurerkellner 2y ago
This is specifically directed at agent traces and not necessarily other LLM use cases. We work on a lot of automated analysis and error detection mechanisms (see https://github.com/invariantlabs-ai/invariant/tree&#
59.
by lbeurerkellner 2y ago
One of the authors here. Let us know what you think :)
60.
Invariant Benchmark Registry: Understanding Agentic Intelligence(explorer.invariantlabs.ai)
1 points by lbeurerkellner 2y ago|0 comments

More ›