service-omni-command-center-configure

Pass

Audited by Gen Agent Trust Hub on Oct 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration were detected in the analyzed scripts or instructions.
  • [COMMAND_EXECUTION]: The orchestration script uses the official Salesforce CLI (sf) and jq to manage metadata. It includes robust checks to prevent execution in production environments and requires the explicit --apply flag for any deployment operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes Salesforce XML metadata, creating an attack surface for indirect instructions. Evidence: (1) Ingestion points: OmniChannel settings XML retrieved via the sf CLI in scripts/configure-and-report.sh. (2) Boundary markers: None. (3) Capability inventory: The skill can deploy metadata changes to the Salesforce org. (4) Sanitization: The Python script settings_document.py validates all modified field names against a strict hardcoded whitelist and ensures all values are valid booleans before rendering the final XML document.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 15, 2026, 07:58 PM