huashu-pentagram-grid
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data, including branding metrics, audience profiles, and strategic messaging, to populate an HTML template. This ingestion of external data creates a surface for indirect prompt injection attacks where malicious input could attempt to subvert agent instructions or inject scripts into the resulting presentation files.\n
- Ingestion points: Data is ingested through user prompts for the deck's topic, audience, source materials, and specific numeric metrics as defined in
SKILL.mdandopen-design.json.\n - Boundary markers: There are no specific instructions for the agent to use XML tags, triple backticks, or other delimiters to isolate user-provided data from its internal operating instructions.\n
- Capability inventory: The skill utilizes file system write capabilities (
fs:write) to produce the outputindex.htmlfile.\n - Sanitization: The instructions do not mandate escaping or sanitization of user-provided strings before they are interpolated into the HTML document structure.
Audit Metadata