agentforce-d360-analyze
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the official Salesforce CLI tool (sf) using the subprocess module to perform organization authentication and display organization metadata. All command-line arguments are either static or derived from inputs validated against strict regular expressions.
- [EXTERNAL_DOWNLOADS]: The skill connects to the Salesforce Data Cloud Query REST API using authorized OAuth tokens to retrieve session audit logs. These network operations are restricted to Salesforce domains.
- [SAFE]: Robust input validation is implemented in the skill's core architecture through fs_guard.py and paths.py. Every path segment, including organization IDs and session UUIDs, is validated against anchored regular expressions to prevent path traversal and filesystem manipulation.
- [SAFE]: SQL queries used for data discovery and fetching are constructed using a centralized escaping helper that doubles single quotes, mitigating potential SQL injection vulnerabilities from user-supplied search terms.
Audit Metadata