I started using Claude Code (claude --dangerously-skip-permissions) and Codex (codex --yolo) and realized I had no reliable way to know what they actually did. The agent's own output tells you a story, but it's the agent's story.
logira records exec, file, and network events at the OS level via eBPF, scoped per run. Events are saved locally in JSONL and SQLite. It ships with default detection rules for credential access, persistence changes, suspicious exec patterns, and more. Observe-only – it never blocks.
https://github.com/melonattacker/logira
But it’s a pain to review..
I suggest adding Stop & PreCompact hooks to your agent which give it the log and ask it to review its own actions in case it did anything unsafe or unexpected. It can check what it sees against what it remembers and tell you if anything stands out.
Or you could give the transcript and log to a new model and have that one do a review.. but either way the goal is to reduce your cognitive load.
Even cooler is when you notice you can have the model provide recommendations - and make its own plan to incorporate them :)
For an example, here’s what I’m doing with transcripts: https://codeleash.dev/docs/self-reflection
reply