html-communication
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill converts user-supplied project details and data into HTML artifacts, creating a potential surface for indirect prompt injection via the source content.
- Ingestion points: User instructions regarding plans, specs, reports, and findings in SKILL.md.
- Boundary markers: No explicit markers are defined to delimit external data from generated markup.
- Capability inventory: The skill generates code (.html, .css, .js) and utilizes the html-upload tool.
- Sanitization: Not explicitly mentioned, though the guidelines focus on semantic structure and clarity.
- [SAFE]: No malicious patterns, obfuscation, or persistence mechanisms were found. The skill includes security-positive requirements such as inlining all CSS/JS and forbidding external network requests in the produced artifacts.
Audit Metadata