Skip to content

Security: PulsarOSDevTeam/pulsarcode

Security

SECURITY.md

Security policy

Thank you for taking the time to make pulsarcode more secure.

Supported versions

We only support the latest release tag on this repository. Issues found in older versions are typically already fixed in the current release; upgrade first, retest, then report what you still see.

Reporting a vulnerability

Please do not open public issues for security disclosures. This repo has Issues disabled to keep the read-only posture clean, and a public security disclosure that names the bug before a fix is shipped puts every user of the latest release at risk.

Email yassine@pulsaros.ca with:

  1. A clear title and one-paragraph summary of the issue.
  2. The minimum reproduction (a few lines of shell or Python).
  3. The version you tested (pulsarcode help or check CHANGELOG.md).
  4. Your OS, Python version, and a rough sketch of the threat model you have in mind.

We acknowledge security reports within five business days. We do not promise a fix on any particular timeline, but we work in good faith toward a coordinated disclosure window after a fix is published.

Out of scope

The following are out of scope for this repository's security policy:

  • Bugs in the upstream Claude Code binary (report those to Anthropic).
  • Bugs in NVIDIA NIM (report those to NVIDIA).
  • Bugs in any model served by NVIDIA NIM (report those to the model vendor).
  • Feature requests or quality-of-life suggestions (we do not accept these on this repository; see CONTRIBUTING.md).

What we consider a security issue here

  • Unauthorized exfiltration of a user's NVIDIA NIM key from ~/.pulsarcode/nim.key.
  • Adapter binding to a non-loopback interface.
  • Outbound network traffic to any host other than integrate.api.nvidia.com (or whatever the user explicitly configured via NVIDIA_NIM_API_BASE).
  • Command injection in the launcher or the adapter.
  • Path traversal in the installer.
  • Privilege escalation through the symlinks the installer creates.
  • Any leakage of prompt content to a third party other than NVIDIA NIM.

What we do not consider a security issue

  • The fact that prompts you send to a model are visible to NVIDIA. That is the documented architecture; if you do not want this, do not use NVIDIA NIM. The local adapter binds 127.0.0.1 only and only forwards to NVIDIA when you make a request.
  • The fact that AGPL-3.0 requires you to disclose modifications if you run a modified version as a network service. That is the license, not a vulnerability.
  • The fact that the launcher writes the NIM key to disk at ~/.pulsarcode/nim.key mode 0600. We chose disk-stored over environment-variable so users can run multiple terminal sessions without re-pasting; if you want stricter handling, fork and modify your fork.

Hall of fame

We list responsible reporters here, by name (with permission) or anonymously, after each fixed vulnerability ships. The list is empty today; once a reported vulnerability lands a fix, the reporter's entry appears here.

There aren't any published security advisories