mom-recall
Warn
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill constructs a shell command by interpolating a user-provided natural language query into
mom recall "<query>". This creates a risk of command injection, as shell metacharacters (e.g.,;,|,&&) in the user input could be used to execute unintended secondary commands within the bash environment. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing untrusted data for use in a privileged tool (Bash).
- Ingestion points: User-supplied search query in
SKILL.mdprocessed at runtime. - Boundary markers: The input is wrapped in double quotes in the Bash command; however, no instructions are provided for the agent to validate or sanitize the input, nor are there boundary markers to distinguish instructions from data.
- Capability inventory: Access to the
Bashtool with a prefix filter formom recallcommands. - Sanitization: No sanitization, escaping, or validation logic is applied to the user-supplied query prior to execution in the shell.
Audit Metadata