memory
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION] (LOW): The skill is susceptible to Indirect Prompt Injection because it stores unvalidated user input for later retrieval and execution.
- Ingestion points: User's casual speech is captured and written to JSONL files in the
.memory/directory via thefs_writeaction. - Boundary markers: The skill lacks explicit delimiters or instructions to the LLM to treat retrieved memory as untrusted data rather than active instructions.
- Capability inventory: The skill utilizes
fs_write(to store),fs_read(to load profile), andfs_grep(to search), creating a feedback loop where stored data is fed back into the agent context. - Sanitization: No sanitization or escaping mechanisms are described for the
contentfield before it is persisted.
Audit Metadata