reqlens

reqlens is an AI/LLM-assisted security proxy and traffic analysis project for request/response fingerprinting, sensitive-value masking, repeated-pattern caching, anomaly detection, and replay-oriented incident investigation.

The repository is currently bootstrapped with project operating guidance in AGENTS.md and safe local configuration defaults in .env.example.

Bootstrap

1. Copy .env.example to an untracked local env file.
2. Fill only local non-secret defaults in that file.
3. Put API keys, bot tokens, webhook secrets, and provider credentials in Forgejo repository secrets or another private secret store.
4. Keep raw traffic captures and replay inputs out of commits.

First implementation boundaries

• Mask sensitive values before storage, replay, or LLM submission.
• Fingerprint canonical request and response structures deterministically.
• Cache repeated LLM analysis results by fingerprint.
• Keep raw capture and replay workflows explicitly opt-in.