service-omni-agent-work-sharing-configure
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the Salesforce CLI (
sf) andjqto manage metadata and query organization information. These operations are essential to the skill's purpose and are limited to the target Salesforce environment provided by the user. - [INDIRECT_PROMPT_INJECTION]: The skill ingests XML metadata from Salesforce organizations. It uses a regex-based extraction method (
sed) to process values for re-deployment. The potential for injection via malicious metadata is mitigated by the strict XML tag matching and a mandatory read-back verification step that validates the resulting state matches the intended configuration. - [DYNAMIC_EXECUTION]: The script generates temporary metadata files (
AgentWork.object-meta.xml) and configuration files (sfdx-project.json) within a restricted temporary directory to facilitate deployments. This is a standard and necessary pattern for Salesforce metadata operations and includes automated cleanup via shell traps. - [SAFE]: The skill implements significant safety guards, specifically checking the
OrganizationTypeandIsSandboxstatus to block unauthorized changes to production environments. It also enforces a dry-run mode (--plan) and requires an explicit flag (--confirm-org-wide-visibility) for any operation that expands record access.
Audit Metadata