platform-metadata-discovery-get
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes metadata content such as field labels and object descriptions from external Salesforce orgs, which constitutes an untrusted data source.
- Ingestion points: Metadata fetched via
sf data query,sf sobject describe,sf org list metadata, and thesearch_metadataMCP tool. - Boundary markers: The skill does not define explicit delimiters or instructions for the agent to isolate external metadata content from its internal reasoning processes.
- Capability inventory: The skill is designed to work with the
sfCLI and MCP tools, which can perform data queries and metadata retrieval. - Sanitization: No sanitization or filtering logic is provided for the external metadata descriptions processed by the skill.
Audit Metadata