Working on https://kapturekafka.dev, a desktop app for Kafka protocol inspection. Think Wireshark or Fiddler, but native for Kafka.
Useful to debug local Kafka apps against any cluster, intercepts the traffic, decodes the protocol. You see interesting (and weird) things when you look at the protocol. Still early, though already useful for local debugging when you know what you want.
I run a lot of small background services on my Mac (thanks Claude/Codex), and Homebrew services that I barely look at. It felt surprisingly hard to know what was actually happening on my machine and investigate/fix quickly, so I built Launchdeck: a fast TUI for launchd jobs and Homebrew services.
- Make the validator read-only for the agent. Mount it as read-only in the container, or hash your eval scripts at startup and verify before each run. If the agent can write to anything in its evaluation path, it can (will) game it.
- Log the full trajectory, not just the output. Every tool call, file diff, reasoning step. Then run a second agent over the trace with no knowledge of the KPI: it only knows what honest execution looks like and will use its internal honest alignment to assess it.
- Write system prompts like job descriptions, not optimization targets. Name a reviewer. Give the agent permission to fail ("if you can't hit the target, explain why").
- Walk your own prompts: what's the metric, what can the agent write, and can it reach the metric by modifying the measurement instead of doing the work? If yes, close that path.
Wanted to build a curated list of truly free, publicly accessible real-time datasets and streaming sources to have real-life data, and see their shape quickly in the browser (WebSocket, SSE). Also led me to build kafka-connect-websocket (https://github.com/conduktor/kafka-connect-websocket) to publish into Kafka directly.
What do you think, what have you saw in terms of usage/adoption/issues?