kxbnb

5 ms·

31.
by kxbnb 9mo ago
This spec reads like what we've been building at toran.sh - transparent proxy for AI API calls with observability. The core idea: you create a "toran" (read-only inspection endpoint) bound to a single upstream. Point your cli
32.
by kxbnb 9mo ago
The escalation logic (lightweight fetch → full browser only when needed) is a nice optimization. That's exactly the kind of thing that's painful to build yourself. Curious about debugging: when an agent's request fails or ret
33.
by kxbnb 9mo ago
The MCP Server integration is a great addition - being able to have Claude manage VMs directly opens up interesting sandboxing patterns for agent workflows. One thing I've been thinking about with agents running in isolated environment
34.
by kxbnb 9mo ago
This really resonates - the opacity problem is exactly what makes MCP-based agents hard to trust in production. You can't control what you can't see. We built toran.sh specifically for this: it lets you watch real API requests fro
35.
by kxbnb 9mo ago
Love the security-first approach with Docker sandboxing - that's often an afterthought in agent wrappers. The BYOK model is compelling too for teams with compliance requirements. Curious: how are you thinking about policy enforcement f
36.
by kxbnb 9mo ago
Love the CLI-first approach for secrets. The Rust implementation should give you solid performance for dev workflows. How does it handle rotation policies and audit logs? We've found that API key rotation is often where teams struggle
37.
by kxbnb 9mo ago
This is a smart approach to API testing - capturing real production patterns is way more valuable than synthetic tests. One question: how do you handle sensitive data in the captured traces? We've been working on API governance at tora
38.
by kxbnb 9mo ago
That matches what we’ve seen too. Reliability seems to come from tightening the scope of each step, not from making the agent itself more sophisticated. The intent boundary is the key thing for us. Rather than reasoning about an agent end t
39.
by kxbnb 9mo ago
Hi HN, I built toran because I kept running into situations where I could not easily tell what my app was actually sending to third party APIs when things went wrong. toran is a read-only outbound API inspector. You swap your API base URL f
40.
Show HN: toran – a read-only outbound API inspector(toran.sh)
1 points by kxbnb 9mo ago|1 comments
41.
by kxbnb 9mo ago
Hi HN, I built Keypost after running into the same problem repeatedly with MCP servers. They expose powerful tools, but there is no standard way to control who can call what, with which parameters, or under what limits. Keypost sits in the
42.
Show HN: Keypost – Policy enforcement for MCP pipelines(keypost.ai)
3 points by kxbnb 9mo ago|2 comments