May 19
Loading video player...
It's getting harder to see what actually ran once execution begins. Most tools capture requests, traces, or logs — not real behavior. In MCP systems, tools call other tools, dependencies execute code indirectly, and subprocesses spawn during normal operation. Without visibility at that layer, teams reconstruct behavior after the fact. BlueRock MCP Python Hooks is an open-source Python sensor that captures runtime activity directly from within a Python application. The hooks attach at interpreter startup, before application code runs — no code changes, no SDKs. What the hooks capture: • MCP protocol activity — tool calls, session lifecycle, client/server connections • Resource access triggered by tools • Module imports across the dependency chain, with version and SHA-256 verification • Process-level activity such as subprocess execution All events are emitted as structured NDJSON and written locally — inspect directly or forward into OpenTelemetry, Grafana, or any pipeline you already run. Apache 2.0. Repo → github.com/bluerock-io/bluerock Quickstart and full docs → bluerock.io/developer-hub/mcp-python-hooks Read the launch post → bluerock.io/blog/introducing-bluerock-mcp-python-hooks