canvas-assignment-design
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user-provided data and use it to generate structured content for Canvas assignments.
- Ingestion points: Phase 1 (Learning Objective Clarification) in
SKILL.mdasks for student objectives, assignment types, and course context, all of which are external data sources. - Boundary markers: The assignment template provided in Phase 3 does not utilize distinct delimiters or specific warnings to prevent the AI from treating embedded text in user inputs as instructions.
- Capability inventory: The skill includes instructions to use
create_assignmentandcreate_rubric(via thecanvas-mcpintegration) which allows it to commit generated content directly to an external platform. - Sanitization: There are no explicit steps to sanitize or filter educator-provided input before it is interpolated into the final assignment description markdown.
Audit Metadata