service-itsm-agentic-setup-incident-sla-configure

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill interacts exclusively with official Salesforce REST and Connect API endpoints via the vendor-provided headless-360 MCP tool. All operations are scoped to the trusted author's (forcedotcom) domain.
  • [SAFE]: The instructions regarding output formatting—such as omitting internal record IDs and masking technical sentinel strings—are consistent with professional data privacy and user experience standards, rather than malicious concealment.
  • [SAFE]: The skill implements a rigorous safety framework, including a prerequisite gate (Phase 0.5) that requires explicit, informed user consent before enabling permanent organization features like SLA Versioning.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a standard indirect prompt injection surface as it ingests data from the Salesforce organization (e.g., Account names and picklist values).
  • Ingestion points: Account names (Phase 1 step 5), Incident picklist values (Phase 1 step 3), and custom milestone names.
  • Boundary markers: The skill uses multiple AskUserQuestion confirmation steps and mandatory plan narration to ensure the user reviews all configuration before changes are applied.
  • Capability inventory: Operations are limited to Salesforce configuration via MCP tools.
  • Sanitization: The skill explicitly describes escaping Account names in SOQL queries (SELECT ... WHERE Name = '<escaped>') to prevent injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 09:40 AM
Security Audit — agent-trust-hub — service-itsm-agentic-setup-incident-sla-configure