claude-landing-composer
Pass
Audited by Gen Agent Trust Hub on Jun 12, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) by processing potentially untrusted data to generate and execute code.
- Ingestion points: The skill processes user-provided landing page requirements and buyer persona data retrieved from the
icp-deep-scannerskill in SKILL.md. - Boundary markers: There are no explicit delimiters or instructions to ignore embedded instructions within the processed data.
- Capability inventory: The skill is authorized to use
Writefor file creation andBashfor shell command execution as seen in the tools configuration. - Sanitization: No sanitization, escaping, or validation of the external content is performed before it is used to generate components or shell commands.
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to execute build and development commands (npm run dev/build) in SKILL.md. This capability allows for the execution of code derived from the aforementioned untrusted inputs.
Audit Metadata