website-build
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
Bashtool to initialize the directory structure for the website and executes a provided Python script (generate-css.py) to process design variables into a CSS stylesheet. - [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests content from
website-project.json,website-plan.json, and narrative source files which are subsequently passed into prompts for page-generation agents. Evidence: 1. Ingestion points: Project configuration JSONs and markdown narrative files. 2. Boundary markers: None implemented to isolate external content from instructions. 3. Capability inventory: File system writes (Write), shell access (Bash), and multi-agent orchestration (Agent). 4. Sanitization: Validation is restricted to identifying unedited«TODO: ...»markers in legal text.
Audit Metadata