platform-value-set-generate
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input (e.g., picklist labels and API names) to generate Salesforce metadata files.
- Ingestion points: User-provided text strings used to define
<fullName>,<label>, and other XML elements inSKILL.md. - Boundary markers: The instructions include explicit validation rules to reject and explain invalid inputs (e.g., Section 6), which helps mitigate accidental obedience to embedded instructions.
- Capability inventory: The skill uses the
sfCLI to query metadata from the Salesforce environment (SKILL.md). - Sanitization: The skill provides logic for sanitizing developer names (e.g., converting spaces to underscores) and checking for uniqueness within a set.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the Salesforce CLI (
sf data query) to verify the existence of standard values in the target environment. This is standard functionality for Salesforce development workflows. - [EXTERNAL_DOWNLOADS]: The skill references official Salesforce documentation (developer.salesforce.com) to provide a list of valid standard value set names. This is an authoritative and expected source for this technical context.
Audit Metadata