service-email-to-case-configure

Pass

Audited by Gen Agent Trust Hub on Oct 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes the Salesforce CLI (sf) using the Python subprocess module to manage authentication, query organization data, and deploy metadata configuration. These calls are well-defined and target standard administrative operations.
  • [DYNAMIC_EXECUTION]: The test bootstrap script (scripts/tests/_bootstrap.py) uses importlib.util to dynamically load Python scripts that have hyphens in their filenames (e.g., apply-casesettings.py). This is a legitimate implementation used solely for unit testing and does not impact runtime execution of the primary skill logic.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external XML data provided by the user to configure Email-to-Case settings.
  • Ingestion points: scripts/apply-casesettings.py reads XML configuration files via the --input argument.
  • Boundary markers: The script uses the xml.etree.ElementTree parser to handle structured data, reducing the risk of the agent misinterpreting data as instructions.
  • Capability inventory: The skill has the capability to execute shell commands (sf CLI) and perform network operations (Salesforce SOAP/REST APIs) as part of its primary function.
  • Sanitization: A dedicated validation script (scripts/validate-casesettings.py) enforces strict structural and value-based rules on the input XML before any mutation occurs.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 15, 2026, 01:29 PM