xaf-web-api

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents the creation of a backend service designed to ingest and process untrusted data from external sources.
  • Ingestion points: Data enters the system via OData query parameters ($filter, $expand), RESTful JSON payloads in POST/PATCH requests, and custom action parameters as shown in SKILL.md.
  • Boundary markers: The instructions rely on standard HTTP request boundaries and XAF's internal ObjectSpace filtering logic.
  • Capability inventory: The skill demonstrates capabilities for persistent data modification (os.CommitChanges()), user authentication, and performing network requests using HttpClient.
  • Sanitization: The documentation highlights the integration of XAF's Security System, emphasizing object-level and member-level permissions as primary defense mechanisms.
  • [DATA_EXFILTRATION]: The skill contains examples of network operations targeting non-whitelisted domains.
  • Evidence: The HttpClient example in SKILL.md demonstrates a network connection to https://myapp.com to perform authentication and data retrieval.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 09:13 AM
Security Audit — agent-trust-hub — xaf-web-api