pitch-landing
Warn
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands using the
Bashtool to invoke thepitch_agent.uiuxPython module. - [COMMAND_EXECUTION]: Potential shell command injection vulnerability in
SKILL.md. The workflow steps interpolate the{product_type}variable directly into apython3command:python3 -m pitch_agent.uiux search --domain landing --query "{product_type}" --limit 3. If the user-supplied input contains shell metacharacters like semicolons or pipes, it could allow arbitrary command execution. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes untrusted user input through search queries.
- Ingestion points: The
{product_type}variable is used as a query parameter inSKILL.md. - Boundary markers: No delimiters or instructions to ignore embedded commands are present when interpolating user data.
- Capability inventory: The skill possesses
Bashexecution capabilities. - Sanitization: There is no evidence of input validation or sanitization before the data is passed to the search tool.
Audit Metadata