service-agentforce-channel-configure
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a mandatory production write-guard in Phase 0, verifying the target organization's type to prevent accidental modifications to production environments.\n- [SAFE]: Critical routing changes are protected by a live-traffic gate in Phase 2, which requires explicit user consent through the AskUserQuestion tool.\n- [SAFE]: All metadata and data operations are conducted using official Salesforce CLI tools and local Python scripts for validation, with no external dependencies or remote code execution.\n- [SAFE]: Data processing is limited to internal Salesforce metadata and records, with no network exfiltration to external domains detected.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests Salesforce metadata, which is handled in a controlled administrative context and validated structurally by local scripts, presenting no significant risk.\n
- Ingestion points: Salesforce metadata and records queried in
SKILL.mdandreferences/channel-branch-email.md.\n - Boundary markers: None explicitly implemented.\n
- Capability inventory: Metadata deployment, data record creation/updates, and shell command execution using standard CLI tools.\n
- Sanitization: Local Python scripts
scripts/validate-botemaildefinition.pyandscripts/validate-emailtemplate.pyperform structural and field-level validation.
Audit Metadata