| Version | Supported |
|---|---|
| 0.4.x | Yes |
| 0.3.x | Yes |
| < 0.3 | No |
Please do not report security vulnerabilities through public GitHub Issues.
If you discover a security vulnerability in @pmatrix/claude-code-monitor, please report it by emailing:
Include the following in your report:
- Description of the vulnerability
- Steps to reproduce
- Potential impact
- Suggested fix (if any)
You will receive a response within 72 hours. We will keep you informed of the progress and notify you when the issue is resolved.
This policy covers:
- The
@pmatrix/claude-code-monitornpm package - The Claude Code hook runtime (Safety Gate, Credential Scanner, Kill Switch)
- The P-MATRIX server API (
api.pmatrix.io)
- Vulnerabilities in Claude Code itself (report to Anthropic)
- The content of agents monitored by this plugin
- Third-party dependencies (report to the respective maintainers)
@pmatrix/claude-code-monitor is designed to be content-agnostic:
- LLM prompts and responses are never transmitted to P-MATRIX servers
- Credential scanning runs entirely on-device
- Pattern-based blocks (
rm -rf,sudo,curl | sh) have no network dependency - Data sharing is opt-in and transmits only numerical behavioral metadata
For full privacy details, see the README.