agent-memory-mcp
Fail
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The installation instructions require cloning a remote repository and executing its contents using
npm installandnpm run compile. - [EXTERNAL_DOWNLOADS]: During setup, the skill fetches an entire repository from an external source at
https://github.com/webzler/agentMemory.git. - [COMMAND_EXECUTION]: The skill requires the execution of multiple shell commands to start an MCP server and a dashboard, which are given access to the local workspace via absolute paths.
- [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection through its memory system.
- Ingestion points: Data enters the agent's context through the
memory_writetool and synchronized project documentation. - Boundary markers: No delimiters are specified to distinguish between stored knowledge and potential instructions.
- Capability inventory: The skill has file system access and command execution capabilities via the MCP server.
- Sanitization: No sanitization or filtering logic is provided for the content stored in or retrieved from the memory bank.
Recommendations
- AI detected serious security threats
Audit Metadata