🍽️ Stop Feeding Raw Code to AI Agents. Give Them a Menu Instead. Sentinel-IR

#agents #ai #javascript #opensource

Release: "sentinel-ir@0.1.2 on npm" (https://www.npmjs.com/package/sentinel-ir/v/0.1.2)
Source: "GitLab repository" (https://gitlab.com/jakubrojicek11/sentinel-ir)
Release changes: "Compare v0.1.1 with commit adf875a1" (https://gitlab.com/jakubrojicek11/sentinel-ir/-/compare/v0.1.1...adf875a1502e3bcf7242bff85a32dcf62f023c07)

Most people building AI agents today make the same expensive mistake: they feed raw, chaotic source code directly into the LLM.
It’s like driving to the supermarket, buying raw ingredients, and cooking for hours every time you want a quick snack. It is incredibly SLOW, expensive, and the AI often gets lost in the recipe.
That’s why I created Sentinel-IR.

What is Sentinel-IR?

Think of it not as an antivirus, but as a Context Extraction Engine.
Instead of forcing your AI Chef to read thousands of lines of code to find security risks or analyze structures, Sentinel-IR pre-processes the code into a clean, structured "Menu."
• The Problem Today: You ask the LLM: "Is this 10,000-line code safe?" The LLM burns through millions of tokens, cooks the server, and takes minutes to answer.
• The Sentinel-IR Way: Our AST analyzer strips out the garbage and delivers a precise, lightweight Intermediate Representation (IR).

How it works at the table:

When the LLM Chef gets the Sentinel-IR Menu, it doesn't need to invent a recipe. It instantly reads the order:
• 🪑 Table 10 wants a strong drink (child_process.execSync).
• 🚫 Special request: But hold the pickles! (The analyzer caught an attacker trying to hide a malicious payload inside a parenthesized_expression).

Why it matters:

By giving the LLM a clean, structured semantic extract instead of raw text, your AI agents make security and logic decisions in milliseconds instead of minutes—saving you a massive fortune on OpenAI token bills.
Sentinel-IR is now live on npm (v0.1.2). Let's make AI agents fast again.

Independent Engineering Projects

Built as independent engineering experiments across AI, software analysis, security, and autonomous systems.

Evaluated by reproducible evidence, not by the number of layers.

Explore the projects:

Sentinel Ecosystem: AutoDoc-Sentinel, Sentinel-IR, Sentinel-IR-MCP and related tools.
Agentix Lite: Independent security and system-observation project.

Find the projects: GitLab repositories | NPM packages

Support the work: ⭐ Star the repositories on GitLab.

Connect: Follow me here on DEV Community or explore the Playground.

Each project is distributed under its own license. Check the corresponding repository or NPM package for licensing terms.