omnistudio-omniscript-generate

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes shell scripts (scripts/check-duplicate-omniscript.sh and scripts/deploy-omniscript.sh) that execute Salesforce CLI (sf) commands. These scripts implement input validation using a regular expression (^[a-zA-Z0-9_\ ]+$) to sanitize user-provided arguments before they are used in shell commands, effectively mitigating typical command injection risks.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it processes user-provided requirements (such as Type, SubType, and Language) and interpolates them into metadata templates and shell script arguments.
  • Ingestion points: Input parameters in scripts/check-duplicate-omniscript.sh and scripts/deploy-omniscript.sh are derived from user-defined requirements.
  • Boundary markers: None identified for the shell script arguments to explicitly separate instructions from data.
  • Capability inventory: The skill performs shell execution, Salesforce metadata queries, and deployments via the sf CLI.
  • Sanitization: Alphanumeric and underscore validation is used in the shell scripts to restrict the characters allowed in input strings.
  • [SAFE]: The skill includes comprehensive security guidance in references/best-practices.md, which explicitly warns against storing sensitive data (like SSNs, passwords, or tokens) in client-side JSON and recommends that sensitive logic be handled server-side.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 01:01 PM
Security Audit — agent-trust-hub — omnistudio-omniscript-generate