canvas-course-qc
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection (Category 8) as it processes untrusted data from Canvas courses. Ingestion points: Untrusted data is retrieved from
get_course_structure,list_assignments, andget_page_content. Boundary markers: The skill lacks explicit markers or instructions to ignore embedded prompts within the course data. Capability inventory: The skill has the ability to modify course content using theupdate_moduleandbulk_update_pagestools. Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the LMS. - [COMMAND_EXECUTION]: The skill uses Canvas MCP tools to perform course audits and updates. All operations that change course state, such as publishing items, require explicit user confirmation, which serves as a security control.
Audit Metadata