Prompt Guard–MitM proxy that blocks secrets before they reach AI APIs
MITM proxy catches leaked secrets before they hit AI APIs — better than post-hoc scanning.

ML-based secret detection stops real leaks before ChatGPT sees them; better accuracy than regex tools.
Developers and teams using GenAI tools (ChatGPT, Claude) who risk leaking credentials and sensitive data
Snyk · git-secrets · TruffleHog
We just launched Kasbah Guard, a free browser extension that detects and blocks 50+ secret formats (API keys, tokens, credentials, AWS keys, GitHub PATs, etc.) using ML entropy scoring + context filtering.
It runs 100% locally in your browser (no server calls, no data collection). We built this because existing tools (like Snyk) only catch ~30 secret types, and regex-based detection has too many false positives.
Key innovation: Naive Bayes multi-feature classifier that analyzes Shannon entropy, character transitions, bigram frequency, and context to distinguish real secrets from test fixtures and docstrings.
Live on Chrome, Firefox, Edge, Opera, Safari. All 5 browsers use identical detector.js (263+ tests passing).
Happy to answer questions about the ML model, false positive suppression, or why we built this.
https://bekasbah.com https://chrome.google.com/webstore/detail/kasbah-guard/...
Would love HN feedback!
MITM proxy catches leaked secrets before they hit AI APIs — better than post-hoc scanning.
Catches AI code smells ESLint misses: empty catches, hardcoded secrets, god functions.
OWASP-backed security layer specifically for AI agent memory poisoning risks.
Yet another secret scanner when TruffleHog and Gitleaks already dominate this space.
OWASP reference implementation for ASI06 memory poisoning with LangChain and AutoGen integrations.
Intercepts pastes, masks 30+ secret types locally—zero network requests or tracking.