sf-datacloud-harmonize
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the Salesforce CLI (sf) with the data360 plugin to manage Data Model Objects (DMOs), identity resolution rules, and field mappings. These commands are executed within the user's authenticated environment to perform legitimate administrative tasks.
- [DYNAMIC_EXECUTION]: The skill's workflow includes running a local JavaScript diagnostics tool (diagnose-org.mjs) located in the shared sf-datacloud skill directory. This script is used to verify environment readiness and is part of the established toolchain for this suite of skills.
- [INDIRECT_PROMPT_INJECTION]: The skill processes structured metadata and tool outputs from external Salesforce orgs, which constitutes an ingestion surface for untrusted data. However, the risk is minimized by the skill's read-focused capabilities and the absence of high-privilege write or exfiltration operations.
- Ingestion points: Data returned from sf data360 CLI commands and the diagnostics script output.
- Boundary markers: No explicit instruction-level delimiters are present.
- Capability inventory: Interaction with Salesforce metadata via shell commands and execution of local Node.js scripts.
- Sanitization: Relies on the standard LLM reasoning framework for handling external tool outputs.
Audit Metadata