make-a-deck
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to ingest untrusted user data, such as presentation text, brand assets, and imagery, and interpolate them into generated HTML slide decks. The instructions lack specific guidance on sanitizing or escaping this user-provided content, which creates a potential surface for indirect prompt injection attacks.
- Ingestion points: User-provided presentation content, brand assets, and images (referenced in SKILL.md instructions).
- Boundary markers: Absent; no specific delimiters or instructions to ignore embedded commands are provided.
- Capability inventory: The skill utilizes the
copy_starter_componenttool and generates HTML, CSS, and JavaScript. - Sanitization: Not specified; the focus is on design and narrative flow rather than data validation.
- [DYNAMIC_EXECUTION]: The skill utilizes dynamic script generation to provide UI functionality. It instructs the agent to generate
<script type="text/babel">tags for an interactive control interface (tweaks-panel.jsx) and to load a local JavaScript component (deck-stage.js) to handle slide navigation and scaling logic. This represents standard dynamic execution patterns for building interactive web-based outputs.
Audit Metadata