deckcraft
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns were identified. The skill acts as a bridge to the DeckCraft application, providing detailed guidance on CLI, MCP, and control channel usage.
- [INDIRECT_PROMPT_INJECTION]: The skill features capabilities to ingest and process untrusted external data in multiple formats.
- Ingestion points: Imports presentation data from .pptx, .md, and .txt files.
- Boundary markers: Documentation explicitly describes the --automation-read-root and --automation-write-root flags used to confine the application's file access to specific directories.
- Capability inventory: Offers broad capabilities including file read/write, image generation, and application state manipulation via MCP and JSON protocols.
- Sanitization: Utilizes format-specific parsers (e.g., for OOXML and Markdown) and platform-level directory confinement for security.
Audit Metadata