lp-proto-gen
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied data to generate HTML prototypes, creating a surface for potential indirect prompt injection attacks.\n
- Ingestion points: Data enters through command-line arguments (e.g., --name, --tagline) and user-defined JSON configuration files in the script scripts/generate_landing_page.py.\n
- Boundary markers: The script uses HTML escaping for body content which provides a structural boundary, but it lacks specific instructions to ignore embedded agent directives.\n
- Capability inventory: The script has the ability to write to the local file system and launch the default system browser via the webbrowser module.\n
- Sanitization: The tool uses html.escape(quote=True) for most user inputs. It also uses a custom css_safe function for theme variables that removes angle brackets, though it does not prevent CSS-based injection via braces or semicolons.\n- [COMMAND_EXECUTION]: The skill utilizes the Python webbrowser module to open the generated HTML file, which results in the execution of a system command to launch the browser.\n- [SAFE]: The skill is implemented entirely using Python standard libraries and does not include network operations, remote code downloads, or other high-risk behaviors.
Audit Metadata