craft-page
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and existing repository documentation (such as
CLAUDE.md,docs/standards/design-tokens.md) to generate HTML and JavaScript artifacts. This creates a surface where malicious instructions embedded in those sources could influence the agent's output or be executed in the viewer's browser. - Ingestion points: User prompts, repository design standards, and system documentation files (SKILL.md § 2).
- Boundary markers: The instructions do not specify clear delimiters or "ignore" warnings for untrusted content when generating the final HTML.
- Capability inventory: The skill is designed to produce self-contained HTML files that include CSS, SVG, and JavaScript (SKILL.md § 3).
- Sanitization: There are no explicit instructions for the agent to sanitize or escape external content before interpolating it into the generated markup or scripts.
Audit Metadata