canva-automation
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- Prompt Injection (LOW): The skill is vulnerable to indirect prompt injection through data processed from the Canva API.
- Ingestion points: External design titles from 'CANVA_LIST_USER_DESIGNS' and brand template names/placeholders from 'CANVA_ACCESS_USER_SPECIFIC_BRAND_TEMPLATES_LIST'.
- Boundary markers: Absent; no delimiters or 'ignore instructions' warnings are used when processing external design or template data.
- Capability inventory: The skill possesses capabilities to create designs ('CANVA_CREATE_CANVA_DESIGN_WITH_OPTIONAL_ASSET'), organize files ('CANVA_POST_FOLDERS'), and export data ('CANVA_CREATE_CANVA_DESIGN_EXPORT_JOB').
- Sanitization: No sanitization or validation of external content is mentioned.
- External Downloads (LOW): The skill setup requires adding a remote MCP server endpoint (https://rube.app/mcp). While this is the primary intended mechanism and the provider (Composio) is known, it introduces a dependency on external infrastructure.
Audit Metadata