I'm seeking any feedback from the community on an open-source tool that I've been working on. It's essentially a tool that lets you create governance guardrails and insights for Claude Code.
Here are some of the features:
* No data leaves your machine.
* Every policy and rule is controlled by you.
* Live dashboard - every decision logged locally
* Budget limits - usage patterns, cost caps, token and action limits
* Full audit trail - every tool call logged with matched rule and reason
* Defense in depth - extendable policy layers: 9+ deterministic evaluators and an optional semantic layer (another LLM) as an evaluator; PII detection, config integrity checks, and injection detection
* Agent memory view - see what your agent sees
* Coverage and alignment - built-in compliance tags; add your own
I'd love any feedback on whether this can be useful for your workflow? What features do you find interesting? What's missing?
Kind regards, Mo