mem-setup
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill is designed to collect and manage sensitive authentication data, specifically sync tokens (prefixed with
cm_) and user UUIDs for the cmem.ai service. - [DATA_EXPOSURE]: Accesses and modifies sensitive configuration files located at
~/.claude-mem/settings.jsonand within the plugin's root directory (config.json). - [COMMAND_EXECUTION]: Instructs the agent to execute a local script using Node.js (
node "${CLAUDE_PLUGIN_ROOT}/scripts/cmem-hook.mjs" status) to verify the setup. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: Accepts raw user-provided strings (tokens, IDs, or connection blurbs) which are processed to extract configuration values.
- Boundary markers: None provided for the user-supplied input data in SKILL.md.
- Capability inventory: File system write access (config.json, settings.json), zip archive creation for plugin packaging, and shell command execution via Node.js in SKILL.md.
- Sanitization: No automated sanitization or schema validation is performed on the user-provided tokens or IDs before they are written to disk.
Audit Metadata