service-itsm-agentic-setup-uel-user-create

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the mcp__headless-360__dispatch and mcp__headless-360__dispatch_readonly tools to perform REST API operations against Salesforce. This is the intended functionality for an administrative automation skill and uses authorized MCP channels.\n- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it ingests untrusted user input (e.g., identity and HR attributes) which is then interpolated into API requests. \n
  • Ingestion points: identity and HR attribute fields defined in SKILL.md.\n
  • Boundary markers: The skill mandates a manual confirmation step (Phase 3, Step 5) before any records are created, providing a human-in-the-loop safety boundary.\n
  • Capability inventory: The skill has the capability to create users, assign permission sets, and create Account/Employee records via the mcp__headless-360__dispatch tool.\n
  • Sanitization: Data is structured into JSON payloads for standard Salesforce REST API calls, reducing the risk of content being interpreted as instructions by the downstream system.\n- [PRIVILEGE_ESCALATION]: The skill requires high-level administrative permissions (ManageUsers, ManageProfilesPermissionsets, CustomizeApplication, AssignPermissionSets) to create records and assign permissions. These requirements are explicitly documented and consistent with the skill's purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 04:35 PM
Security Audit — agent-trust-hub — service-itsm-agentic-setup-uel-user-create