What is an AI posture analyzer?

An AI posture analyzer scans a codebase to inventory everything AI-related — SDKs, models, agent and MCP configs, and prompt surfaces — and evaluates its static security risk, such as hardcoded API keys, over-permissioned agents, and prompt injection exposure. It produces a ranked list of findings and a score you can verify by hand.

Is the AI Posture Analyzer free, and does it require signup?

Yes, it is free. Public GitHub repositories analyze instantly with no signup. To analyze a private repository, you sign in with GitHub so the scan runs with your own access — nothing from a private repo is stored, and analysis is ephemeral.

What does the analyzer look for in a repository?

It detects AI SDKs and models, agent and MCP configurations, CI usage of AI, CLAUDE.md and agent instruction files, hardcoded API keys and secrets, and prompt injection surfaces. Findings are ranked by severity and rolled into a transparent 0-100 posture score.

AI Posture Analyzer

Paste a GitHub repo and get an inventory of everything AI in it: SDKs, models, agent and MCP configs, CI usage, prompt surfaces. Plus a ranked list of static risks and a score you can check by hand. Public repos analyze instantly; private repos ask you to sign in with GitHub first.

Try:

Working on something private? Analyze it with your own GitHub access.

A static scan is a snapshot. Agents act at runtime.

This report shows what the code and configs say today. Prismor governs what agents actually do: it intercepts every tool call and enforces policy at the moment of action, in production. We log lightweight, anonymized usage stats (repo name, score, issue counts) to understand how the tool is used.

pip install prismor See the AI agent control plane

AI Posture Analyzer

Try:

Working on something private? Analyze it with your own GitHub access.

A static scan is a snapshot. Agents act at runtime.

pip install prismor See the AI agent control plane