activating-datacloud
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands using the
sf data360CLI plugin to perform operations on Salesforce Data Cloud, such as listing platforms, creating activation targets, and managing data actions. - [COMMAND_EXECUTION]: A local Node.js script (
diagnose-org.mjs) is executed to verify the readiness of the Salesforce organization before performing mutations. The script is located within the agent's local skill storage path. - [PROMPT_INJECTION]: The skill ingests configuration data from external JSON files (
target.json,activation.json,action.json) which serves as a potential surface for indirect prompt injection. (1) Ingestion points: Resource creation commands (sf data360 ... -f <file.json>) that read user-provided or generated files. (2) Boundary markers: None. The instructions do not specify the use of delimiters or warnings for these inputs. (3) Capability inventory: File system read access, shell command execution for cloud resource mutation, and local script execution. (4) Sanitization: No explicit sanitization or validation logic is described in the skill instructions.
Audit Metadata