landing-page-generator
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection vulnerability via untrusted data ingestion.\n
- Ingestion points: The
SKILL.mdfile (Step 2) directs the agent to ingest "existing brand content" like blog posts or marketing materials provided by the user to determine a voice profile.\n - Boundary markers: There are no instructions or delimiters specified to ensure the agent ignores instructions potentially embedded within this user-supplied content.\n
- Capability inventory: The skill can generate Next.js/React code and uses a local script (
scripts/landing_page_scaffolder.py) to output files. It also attempts to execute an external script (marketing-skill/content-production/scripts/brand_voice_analyzer.py) to process the brand content.\n - Sanitization: The process for establishing a brand voice profile does not include sanitization or validation of the input content, allowing potential instruction leakage into the generation phase.
Audit Metadata