If you discover a security issue in llm-wiki, please report it responsibly.

Do NOT open a public GitHub issue for security vulnerabilities.

Instead, email: mehmetgoekce@memotech.ch

Include:

• Description of the vulnerability
• Steps to reproduce
• Potential impact

You will receive a response within 72 hours.

Security issues in llm-wiki typically involve:

• Credential leaks: Wiki pages accidentally containing API tokens, passwords, or secrets (the wiki is git-tracked and credentials must stay in L1 memory)
• Setup script: Path injection or file overwrite vulnerabilities in setup.sh
• Template injection: Malicious content in wiki pages affecting template rendering

llm-wiki is designed with security boundaries:

• L1 (memory) is git-excluded — credentials stay here
• L2 (wiki) is git-tracked — no secrets allowed
• Lint rule 6 scans for credential patterns automatically
• Append-only semantics prevent accidental data loss