onenote-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were identified. The skill is designed with a strong security posture, requiring human intervention and semantic verification for all external interactions.
- [EXTERNAL_DOWNLOADS]: The skill references official Microsoft documentation for the Graph API and OneNote service to ensure compliance with current API contracts (references/official-docs.md).
- [INDIRECT_PROMPT_INJECTION]: The skill architecture handles external content from OneNote, creating a potential surface for indirect instructions. Ingestion points: The workflow involves reading OneNote page content to resolve stable targets and perform read-back verification (SKILL.md). Boundary markers: Risks are mitigated through mandatory 'semantic previews' and 'content-owner approval' for every create or update action (SKILL.md). Capability inventory: The skill utilizes standard file management tools (Read, Glob, Grep, Write, Edit) to maintain repository artifacts (SKILL.md). Sanitization: The workflow specifies offline validation of HTML elements, request sizes, and idempotency behavior (SKILL.md).
Audit Metadata