field-service-data-capture-form-deployer-configure

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes JSON specifications from potentially untrusted sources to generate Salesforce Flow metadata, which is then deployed via Tooling API calls. Evidence chain: (1) Ingestion point: Input JSON specification described in the SKILL.md input contract. (2) Boundary markers: Absent; the JSON content is parsed directly into metadata structures. (3) Capability inventory: Access to the execute_api tool for performing REST and Tooling API operations against a Salesforce organization. (4) Sanitization: PascalCase validation for Flow API names and structured mapping to a verified JSON schema.
  • [PRIVILEGE_ESCALATION]: The skill includes instructions to modify the security posture of the connected Salesforce organization. Specifically, it guides the modification of Org-Wide Defaults (OWD) for DynamicDataCapture and WorkPlan objects to Public Read/Write and patches FieldServiceSettings metadata. These changes are intended to facilitate data visibility for mobile users but represent administrative-level security configuration changes.
  • [SAFE]: All network operations are performed against the authorized Salesforce organization's APIs using standard platform tools. No evidence of unauthorized external network exfiltration, local shell command execution, or access to sensitive local environment files was found.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 02:30 PM
Security Audit — agent-trust-hub — field-service-data-capture-form-deployer-configure