mem0-oss-to-platform
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands like
python -candgrepto audit project files and verify the signatures of the installed Mem0 SDK. These operations are restricted to the migration planning phase and target the developer's local environment for technical discovery. - [EXTERNAL_DOWNLOADS]: The skill manages the
mem0aipackage through standard package managers (pip, npm) and references official vendor documentation and dashboard URLs at mem0.ai and app.mem0.ai. These interactions are expected for a migration tool from the official service provider. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes untrusted project code and configuration files. This is mitigated by a structured workflow that prevents autonomous execution.
- Ingestion points: Project source code (Python, TypeScript), dependency manifests (requirements.txt, package.json), and environment configuration files (.env).
- Boundary markers: While explicit prompt delimiters are not used, the skill implements a mandatory human-in-the-loop checkpoint where a migration plan must be reviewed and approved before execution.
- Capability inventory: File system read/write access, shell command execution for inspection and testing, and dependency management capabilities.
- Sanitization: No specific content filtering is performed on ingested code, but the reviewable
MEM0_MIGRATION_PLAN.mdallows users to audit intended changes before they occur.
Audit Metadata