metadata-completeness-auditor
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a read-only workflow for auditing metadata, ensuring no modifications are made to the source data. It also mandates that generated HTML reports must be self-contained with no scripts or external dependencies, which prevents cross-site scripting (XSS) and data exfiltration through the report file.
- [PROMPT_INJECTION]: The skill processes untrusted metadata from SharePoint list items which could contain malicious text patterns designed to influence agent behavior.
- Ingestion points: Item labels and column values are read from the target library in SKILL.md (Step 3).
- Boundary markers: None specified to encapsulate item data during processing.
- Capability inventory: The skill utilizes list_items for data retrieval and can create folders and files on the SharePoint site (Step 6).
- Sanitization: The skill explicitly restricts the HTML output to inline CSS and prohibits scripts (Step 5), mitigating the risk of executing injected content in the report viewer's browser.
Audit Metadata