landing-gen
Warn
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the user to run the
npx ai-landingcommand to generate project documentation. - [EXTERNAL_DOWNLOADS]: Execution of
npxtriggers a download of theai-landingpackage from the public npm registry. - [REMOTE_CODE_EXECUTION]: Running a third-party CLI tool from an unverified provider (LXGIC Studios) allows for the execution of remote code on the local machine.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) by processing project metadata.
- Ingestion points: Reads project name, description, and keywords from
package.json(documented in SKILL.md). - Boundary markers: No delimiters or "ignore embedded instructions" warnings are used when processing external metadata.
- Capability inventory: The tool generates HTML files (
landing.html) and requires shell execution privileges vianpx. - Sanitization: No documentation suggests that the input metadata from
package.jsonis validated or sanitized before being used in the AI generation process.
Audit Metadata