canvas-automation
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill connects to the official Rube MCP server at https://rube.app/mcp to facilitate Canvas automation.
- [INDIRECT_PROMPT_INJECTION]: The skill's architecture creates a vulnerability surface where instructions embedded in processed Canvas data could potentially influence agent logic.
- Ingestion points: The agent ingests data from Canvas tool outputs and tool schemas via RUBE_SEARCH_TOOLS and RUBE_MULTI_EXECUTE_TOOL.
- Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are included in the skill's workflow to isolate external data from instructions.
- Capability inventory: The skill utilizes capabilities such as RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH, which are capable of modifying data within the Canvas environment.
- Sanitization: There are no defined steps for sanitizing, filtering, or validating external content before it is processed by the agent.
Audit Metadata