Skip to content

Security: pyxis3-ai/.github

Security

SECURITY.md

Security Policy

Reporting a vulnerability

If you discover a security issue in any PYXIS3 repository, please do not open a public issue. Instead, email ops@pyxis3.ai with:

  • A description of the issue
  • A reproducer (proof-of-concept code, exploitation steps)
  • Your assessment of impact

We aim to acknowledge within 48 hours and propose a fix or mitigation within 14 days for in-scope issues.

Scope

In scope:

  • Authentication / authorization bypass in lens (the only PYXIS3 project that hosts a web surface)
  • Code-injection vulnerabilities in vllm-bench (the only one that runs user-supplied benchmarks)
  • Supply-chain issues — typo-squatted dependencies, malicious upstream packages

Out of scope:

  • Issues in upstream projects (vLLM, TGI, Triton, etc.) — please report those upstream
  • DoS attacks against publicly hosted demos (noor.pyxis3.ai, lens.pyxis3.ai) — these are demo instances, not production
  • Issues in documentation repos (pyxis-arch, awesome-model-agnostic-llm, llm-serving-cookbook)

Disclosure

We follow coordinated disclosure. After a fix is released, we credit the reporter (with permission) in the release notes.

There aren't any published security advisories