sf-apex
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed to improve the security posture of Salesforce Apex code by enforcing bulkification, CRUD/FLS checks, and sharing rules.
- [PROMPT_INJECTION]: Instructions specifically task the agent with identifying SOQL injection risks in code, such as string concatenation in dynamic queries, effectively acting as a security linter.
- [DATA_EXFILTRATION]: Adheres to security best practices by recommending Named Credentials for external integrations, preventing the exposure of hardcoded API keys or secrets.
- [COMMAND_EXECUTION]: Uses the standard Salesforce CLI ('sf') for legitimate project deployment and testing operations as specified in the configuration.
- [INDIRECT_PROMPT_INJECTION]: The skill processes local .cls files (Ingestion: Read/Glob tools). While these files could contain malicious comments, the capability scope is limited to Salesforce CLI operations and local file edits, and the agent is instructed to focus on structural code patterns. Boundary markers and input sanitization for the agent's internal prompt are not explicitly defined, but the risk is consistent with standard code-review tasks.
Audit Metadata