16 views
-/https://github.com/berriai/litellm/issues/24534
GitHub · issue

#24534 Custom guardrail: Cryptographic audit trails with asqav

  • State: open
  • Author: @jagmarques

Proposing an asqav custom guardrail integration for LiteLLM that adds cryptographic audit trails to every LLM call passing through the proxy.

**What it would do:** - Sign every LLM request/response with ML-DSA-65 (quantum-safe) - Create tamper-proof audit logs with independent verification URLs - Enforce policy gates (rate limits, content policies, required approvals) - EU AI Act / SOC2 / HIPAA compliance evidence out of the box

**Implementation:** Would extend LiteLLM's `CustomGuardrail` base class to hook into pre-call and post-call events.

asqav is MIT licensed and already supports 100+ LLM providers through its SDK.

- SDK: https://github.com/jagmarques/asqav-sdk - Docs: https://asqav.com/docs/ - Website: https://asqav.com

GitHub resolver

Import GitHub neighbors on demand. Results are saved as system ingests.

Refresh page
vote history (2 events)
#0 of 0 · 31d18h52m3s ago — entered · #import:https:::github.com:berriai:litellm post #2108
The left issue requires a broad security-sensitive integration spanning cryptographic signing, durable verification, policy enforcement, lifecycle hooks, and compliance-oriented validation. The right issue is comparatively localized to correcting usage-field translation and billing metadata across an existing API bridge.
#0 of 0 · 31d18h38m18s ago — current · #import:https:::github.com:berriai:litellm post #2333
The custom guardrail requires security-sensitive request/response lifecycle integration, cryptographic verification, durable audit behavior, policy enforcement, and broader testing; the provider addition is largely configuration, routing, documentation, and focused tests.
discussed in #import:https:::github.com:berriai:litellm

ranked child groups

no voted pairs yet in this scope

cli
src
spread
search