Afif Maahi (GitHub: Kiyoraka) is the security maintainer for AI MemoryCore and the point of contact for vulnerability reports. He reviews every report personally.
Please do not open a public issue for a security problem.
- Use GitHub's private vulnerability reporting on this repository (Security tab, "Report a vulnerability"), or
- Email wings.dark7@gmail.com with
MemoryCore securityin the subject line.
You will receive an acknowledgement within 7 days. Once a report is confirmed, a fix lands on main and the reporter is credited in the commit unless they ask otherwise.
| Branch / snapshot | Supported |
|---|---|
main |
Yes |
| Older snapshots | No |
AI MemoryCore is a file-based memory architecture for AI assistants: Markdown memory files, a master index, a setup guide and a setup wizard. Reports in scope include:
- prompt-injection paths through memory files, diary entries or library items that an assistant loads as context;
- secret or credential leakage in the templates, examples or setup wizard;
- unsafe file handling in the setup flow (path traversal, overwriting files outside the memory folder);
- any instruction in the templates that would make an assistant act unsafely on a user's machine.
Out of scope: vulnerabilities in the AI models or hosting platforms the architecture is used with, and issues in forks that changed the templates.
This project is a template others build companions on. Please test only against your own copy.