Some requests never reach the model
gate.ts short-circuits allowlisted paths/IPs and cryptographically verified clients. Their zero-risk records are local policy outcomes, not model confidence measurements.
Inspect the supporting source ↗REPOSITORY ANALYSIS
traffic-guard
A Node/Python HTTP request gate combining local allowlists and verification with optional TypeSafe questions about bot behavior, attacks, spoofing, traffic category and risk.
E2 · source reviewedPublic material or source evidence was reviewed. A fixed-version source review requires checking a pinned revision; this is not a reproduction, performance endorsement, or safety guarantee.
- Stars
- 1
- Forks
- 0
- License
- —
- Last commit
- 2026-09-21
What Jev does here
traffic-guard progressively evaluates requests against TypeSafe System One using Noul (`is_bot`, `is_attack`, `is_spoofed`), Choice (`traffic_type`), and Score (`risk_level`).
Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.
Pinned gate and policy code reviewed. Short-circuit allow decisions are local outcomes; configured API-outage behavior can differ, and some missing-answer fields receive defaults. No WAF, throughput or bot-detection benchmark was reproduced here.
What task does this project fit?
Combine deterministic request screening and a model-based HTTP traffic assessment.
Material analysed：
From input to action
- Apply path/IP allowlists and local verification checks.
- Ask TypeSafe for bot, attack, spoofing, category and risk judgments where needed.
- Map assessments through the configured allow/challenge/block policy.
Implementation details that matter
Unavailable and incomplete assessments need separate tests
The gate exposes allow/block fallback options and assigns low-risk defaults for some missing answer fields. A deployment must test these paths explicitly instead of treating every returned assessment as a full model result.
Inspect the supporting source ↗When to choose it
Useful for studying staged request gates. Put rate limits and protocol validation ahead of probabilistic screening, and choose an outage policy that matches the application's risk.
How to inspect it yourself
Trace gate.ts's short-circuits, answer defaults and fallback catch, then compare policy.ts for good-bot and honeypot branches using synthetic HTTP metadata.
Limits of this review
- No traffic load or attack benchmark was run.
- The existence of TypeSafe questions does not establish WAF-grade protection or calibrated bot detection.