service-itsm-agentic-setup-cmdb-configure
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external Salesforce API responses, such as tenant provisioning status and feature enablement states, which constitutes a potential attack surface for indirect prompt injection.
- Ingestion points: The skill ingests data from endpoints including
/services/data/v67.0/connect/tenantProvisioningStatusand/services/data/v67.0/connect/setup/discovery/feature/service-cloud-itsm-cmdb-integration/status(File:SKILL.md). - Capability inventory: The skill possesses the capability to modify the Salesforce org state using the
mcp__headless-360__dispatchtool for triggering provisioning and enabling features (File:SKILL.md,references/mcp-invocation.md). - Sanitization: The skill explicitly identifies and mitigates risks by unescaping HTML entities (e.g.,
<,>,",&) and stripping markup fromfailureReasonormessagefields in API responses before displaying them to the user (File:references/mcp-invocation.md). - Boundary markers: The skill requires explicit user confirmation before any write operation and restricts its operations to a specified, trusted MCP toolset.
Audit Metadata