building-omnistudio-integration-procedure
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill promotes high security standards by including a scoring rubric that rewards input validation and the use of Named Credentials while penalizing hardcoded IDs and secrets.
- [COMMAND_EXECUTION]: The
scripts/cli-commands.shfile contains legitimate Salesforce CLI (sf) commands for querying and deploying metadata, which are essential for its purpose as a development tool. - [DATA_EXFILTRATION]: While the skill documents elements for making external HTTP calls, it emphasizes secure practices such as using Named Credentials and avoids any patterns indicative of unauthorized data harvesting.
- [PROMPT_INJECTION]: The skill's instructions focus entirely on technical implementation and do not contain any attempt to bypass AI safety guidelines or manipulate agent behavior.
Audit Metadata