host-memory
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external Markdown files that could potentially contain malicious instructions if the memory store is tampered with by other processes or tools.
- Ingestion points: Reads from
~/host-memory/INDEX.md,~/host-memory/rules/<slug>.md, and~/host-memory/logs/<slug>.md(SKILL.md). - Boundary markers: The instructions do not define specific delimiters for separating user-provided memory content from agent instructions.
- Capability inventory: The skill relies on file system read/write operations to manage the memory structure.
- Sanitization: The instructions include an explicit safety policy: "不记密钥 / token / 密码明文" (Do not record keys / tokens / plain text passwords).
- [COMMAND_EXECUTION]: The skill requires the agent to perform file system operations (read/write/create) to maintain the memory hierarchy in the user's home directory, which is consistent with its stated purpose.
Audit Metadata