shared-memory-intake
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local Python script located at
~/.agents/skills/shared-memory-intake/scripts/validate_submission.pyto perform schema validation and secret scanning on memory submissions. - [PROMPT_INJECTION]: The skill's core functionality involves ingesting data from external files into the agent's shared memory, which creates a surface for indirect prompt injection.
- Ingestion points: Data is read from Markdown files in the
~/.agents/shared-memory-inbox/pending/directory and consolidated intoMEMORY.md. - Boundary markers: The skill does not define specific markers or guardrails to prevent the agent from interpreting data within these files as new instructions.
- Capability inventory: The agent has the capability to read/write files in specific local directories and execute a bundled validation script.
- Sanitization: While a validation script checks for common secret patterns (e.g., private keys, API tokens), it does not include sanitization or filtering to detect or block malicious instructions hidden within the memory content.
Audit Metadata