unpublished-content-report
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes SharePoint metadata, such as file names, titles, and editor names, which are influenced by end-users. This creates a surface for indirect prompt injection if an attacker embeds malicious instructions in file properties.
- Ingestion points: The skill retrieves internal fields and page metadata from SharePoint document libraries and Site Pages (SKILL.md, Steps 4 and 5).
- Boundary markers: The skill instructions specifically require escaping user-controlled values before inserting them into the raw HTML report (SKILL.md, Step 11).
- Capability inventory: The agent has the ability to discover site lists, create new document libraries, and save HTML files to SharePoint (SKILL.md, Steps 2, 3, and 12).
- Sanitization: The instructions include a mandatory step to escape all user-influenced values to prevent cross-site scripting (XSS) or other injection attacks in the generated report.
Audit Metadata