| Version | Supported |
|---|---|
| 0.1.x | Yes |
If you discover a security vulnerability in Jarvis-CLI, please report it responsibly.
Do NOT open a public issue for security vulnerabilities.
Instead, please send a detailed report to the repository maintainers via GitHub private vulnerability reporting:
- Go to the Security tab of this repository
- Click "Report a vulnerability"
- Provide a detailed description of the vulnerability
- Description of the vulnerability
- Steps to reproduce
- Potential impact
- Suggested fix (if any)
- Acknowledgment within 48 hours
- A fix or mitigation plan within 7 days for critical issues
- Credit in the release notes (unless you prefer to remain anonymous)
- Audio is recorded only when explicitly triggered by the user (hotkey press)
- Audio data is processed locally — no data is sent to external servers
- Transcriptions are stored in the system temp directory and marked as consumed after use
- Transcription files use atomic writes to prevent partial reads
- PID files are stored in the system temp directory
- The Claude Code hook only reads local JSON files
- All speech-to-text processing happens locally via faster-whisper
- No network requests are made during normal operation
- The Whisper model is downloaded once from Hugging Face Hub on first use