skills/sfrmrc/skills/corner-lasers/Gen Agent Trust Hub

corner-lasers

Warn

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [OBFUSCATION]: The file demo/index.html contains a large Base64-encoded string assigned to the variable encodedHtml. This string encodes the full HTML and JavaScript source for the skill's visual demo, which is decoded and executed at runtime via an iframe.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to interpret design requirements from user prompts to generate WebGL compositions. This architecture creates an indirect prompt injection surface. 1. Ingestion points: User-provided design descriptions in demo/PROMPT.md and runtime inputs. 2. Boundary markers: The skill instructions do not specify delimiters or instructions to ignore embedded commands in the design brief. 3. Capability inventory: The skill generates and executes HTML/JavaScript/WebGL content in a browser environment. 4. Sanitization: No specific sanitization or validation of the input prompt is documented.
  • [UNVERIFIABLE_DEPENDENCIES]: The skill demo downloads and executes JavaScript libraries from well-known services including Three.js, GSAP, and Iconify. These dependencies are specified with subresource integrity hashes in the demo/source.json metadata.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 26, 2026, 01:37 PM
Security Audit — agent-trust-hub — corner-lasers