carousels

Pass

Audited by Gen Agent Trust Hub on May 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates the execution of internal Node.js scripts (generate-image.mjs and render.mjs) to automate image generation and document rendering tasks. These operations are necessary for the skill's primary functionality and are performed within the user's local workspace.\n- [EXTERNAL_DOWNLOADS]: The skill depends on the 'puppeteer' package and requires the download of browser binaries for slide rendering. These are standard, well-known dependencies for headless browser automation and are documented for the user.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests untrusted source text to generate slide content and prompts. \n
  • Ingestion points: Pasted text or local files ingested during Phase 1 of the skill workflow.\n
  • Boundary markers: The skill does not explicitly use delimiters to separate user content from its own instructions.\n
  • Capability inventory: The agent can write to the local file system and execute internal scripts for image and document generation.\n
  • Sanitization: Source content is used directly for editorial distillation and prompt generation without prior sanitization or filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
May 26, 2026, 10:17 AM
Security Audit — agent-trust-hub — carousels