stacked-data-cards
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill generates standalone, single-file HTML animations. It uses native web technologies such as CSS Transitions and SVG for visual components, and requestAnimationFrame for timeline management, ensuring the generated output remains independent of external scripts or CDNs.
- [SAFE]: Audio functionality is implemented through local synthesis via the Web Audio API. The code includes appropriate error handling for browser autoplay policies and does not require external audio assets.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data to generate visual narratives, creating an attack surface for indirect prompt injection if untrusted external data is processed by the agent. 1. Ingestion points: User-provided arguments, narrative text, or external documents used to populate card templates defined in SKILL.md and assets/template.html. 2. Boundary markers: Not explicitly defined in the templates; the skill relies on the agent's ability to map data correctly into UI components. 3. Capability inventory: JavaScript execution within the generated HTML file, limited to animation logic, timeline control, and local audio synthesis. 4. Sanitization: The skill relies on the agent's default mapping logic; no custom sanitization or validation routines are implemented in the skill scripts.
Audit Metadata