pipeline-review
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill facilitates legitimate Salesforce CRM data analysis. It uses authorized platform tools (
dispatch_readonly,dispatch) to interact with standard Salesforce API endpoints (/services/data/v66.0/graphql,/services/data/v63.0/query). All API interactions are scoped to the authenticated user or a specifically resolved representative, preventing unauthorized data access. - [SAFE]: The skill contains defensive instructions that strengthen the execution environment. It explicitly forbids the agent from using shell commands (e.g.,
cat,jq,python3) on temporary files containing API responses, directing it to use the platform's secure 'Read' tool instead. It also restricts the agent to the documented API paths, prohibiting unnecessary discovery or external endpoint guessing. - [SAFE]: The ingestion of Salesforce Opportunity fields (Name, Amount, NextStep) for reporting is the primary intended purpose of the skill. While this creates a standard attack surface for indirect prompt injection, the risk is mitigated by the skill's use of a structured JSON widget for output and instructions to cite queried values exactly, maintaining data integrity and reducing the potential for malicious instruction execution.
Audit Metadata