M1K3's whole promise is "Your AI. Your Mac. Nothing leaves." — so privacy and security reports are the most valuable contributions this project can receive.
Please do not open a public issue for security problems.
- Preferred: GitHub private vulnerability reporting (Security tab → "Report a vulnerability").
- Or email kevin@round-tower.ie with
[M1K3 SECURITY]in the subject.
You'll get an acknowledgement within 72 hours and a status update within 14 days. If the report is valid we'll credit you in the fix's release notes (or keep you anonymous — your call).
Anything that breaks the local-only promise ranks above a classic RCE for this project:
- Data leaving the machine without explicit user consent (network calls beyond the opt-in web search / model downloads).
- Prompt-injection paths that exfiltrate knowledge-base or memory content through the MCP server or web tools.
- Sandbox or entitlement escapes in the Mac app.
- The local MCP server (
127.0.0.1:4242) being reachable off-host or abusable by other local processes beyond its design. - PII surviving the diagnostic redaction in issue reports.
| Surface | Status |
|---|---|
macOS app (macos/, TestFlight beta) |
Supported — latest beta build |
Python CLI (attic/) |
Archived — best effort only |
間 AI mobile (app/) |
Pre-release — not yet supported |