Sentinel-IR in 30 seconds: what's new in 0.4

description: "Python, native-code detection, an audit log for your AI agent, a browser playground, and now on GitHub, Glama and the official MCP Registry."

series: "Sentinel-IR: what's new"

Sentinel-IR turns a source file into a short structural note for an AI
agent: what the file does, how risky it is, and on which line. The agent reads
the note instead of the whole file. Since our last post:

• sentinel-ir 0.4.5 reads Python too (besides JavaScript and TypeScript), flags native code (ctypes, process.dlopen, FFI) as critical, analyzes a 1 MiB JavaScript file in about 5 s instead of about 170 s, and tells you where broken code stops parsing.
• sentinel-ir-mcp 0.4.6 lets you exclude folders the agent must not read and keep an audit log of what it asked for, with the SHA-256 of every file it was shown. Never your code.
• The playground shows all of it in your browser. Nothing to install, and your code stays on your machine.

New homes: the repos are mirrored on GitHub, and the MCP server is listed
on Glama and in the official MCP Registry.

Your agent now has a diary. You're the only one who can read it.

• Playground: https://sentinel-ir-playground-dc30c8.gitlab.io/
• GitHub: sentinel-ir · sentinel-ir-mcp · sentinel-ir-playground
• Glama: https://glama.ai/mcp/servers/jakcmencz/sentinel-ir-mcp
• MCP Registry: io.github.jakcmencz/sentinel-ir-mcp
• npm: sentinel-ir · sentinel-ir-mcp