knurl / secret & config scanner
Ask an AI to check a config for leaked secrets and it guesses. A scanner doesn't.
Paste a .env, docker-compose.yml, nginx config, or any blob — get a
severity-ranked report of what leaked, where, why it's risky, and how to fix it.
Awaiting scan — paste a config and press Scan.
AWS/GCP/Azure, Stripe, GitHub/GitLab, OpenAI/Anthropic/Gemini/Hugging Face/Groq/Replicate, private keys, JWTs, database URIs, and high-entropy strings.
Debug mode on, 0.0.0.0 binds, disabled/weak TLS verification, privileged containers, default and weak passwords.
Curated, high-confidence rules. Placeholders, UUIDs, git SHAs, and hashes are filtered so the result is trustworthy.
The same scanner runs as a connector — ask "any leaked secrets or misconfigs here?" right in your chat.
Add the knurl Secret Scanner app, then paste & ask. (Directory link coming at launch.)
Add the MCP server https://mcp.knurl.tools/mcp in your client. Setup ships with the connector.
Billed on our own checkout — your account and receipts stay with knurl, not the assistant platform.
No. On this page the scan runs entirely in your browser using WebAssembly-free JavaScript — there is no server call. You can verify it in your browser's Network tab.
Yes. It's fixed rules plus entropy analysis, not a language model. The same input always produces the same findings.
An LLM reads text linearly and can miss or hallucinate secrets. A scanner applies exhaustive rules every time. The knurl connector gives the assistant a real scanner to call instead of guessing.
You can — it never leaves your browser — but rotate anything the scanner flags. Treat a flagged secret as already exposed.