canva
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data through its "Bulk Create" automation feature to generate large volumes of visual content.
- Ingestion points: The skill ingests data from CSV or Excel spreadsheets, including text fields and image URLs, which are used to drive the automated creation of design variants (
references/workflows-and-templates.md). - Boundary markers: The skill incorporates a mandatory human approval step for every visual produced and a manual QA pass for layout issues before publishing (
SKILL.md,references/workflows-and-templates.md). - Capability inventory: When connected, the agent has the capability to drive design creation, resizing, and adaptation through the Canva MCP server and Connect APIs (
SKILL.md). - Sanitization: The instructions focus on manual review of the visual output rather than programmatic sanitization of the input spreadsheet data.
Audit Metadata