skills/mukul975/anthropic-cybersecurity-skills/implementing-data-loss-prevention-with-microsoft-purview/Gen Agent Trust Hub
implementing-data-loss-prevention-with-microsoft-purview
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves data from the Microsoft Graph Security API including DLP alert titles and descriptions. This external content could contain malicious instructions designed to influence the agent when it summarizes or analyzes the reports.
- Ingestion points: The
scripts/agent.pyscript fetches external data through theget_dlp_alertsandget_sensitivity_labelsfunctions. - Boundary markers: The skill lacks specific delimiters or 'ignore' instructions when interpolating external alert data into reports or agent context.
- Capability inventory: The skill has capabilities for file system writing (
csv.DictWriter,Path.write_text), network operations to Microsoft Graph, and administrative PowerShell actions for policy management. - Sanitization: There is no evidence of filtering or sanitization of text fields within the retrieved alerts to prevent interpretation of embedded instructions.
Audit Metadata