sales-agentforce-pipeline-management-configure
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a series of bash scripts to automate Salesforce administration tasks via the Salesforce CLI (sf), curl, and python3. These operations include enabling org settings through SOAP and REST APIs, deploying metadata for Flows and GenAiPromptTemplates, and managing user permission assignments. All commands are executed within the context of the user's authenticated session to Salesforce's well-known domains.
- [PROMPT_INJECTION]: The skill processes external Opportunity data (notes, emails, transcripts) within AI prompt templates, presenting an indirect prompt injection surface. 1. Ingestion points: Untrusted data enters the agent context through scripts/shared/test-opp.sh and scripts/add-field-suggestion.sh via seeding tasks. 2. Boundary markers: Prompt templates in assets/field-completion-template.genAiPromptTemplate-meta.xml use delimiters (XML tags and triple quotes). 3. Capability inventory: The skill possesses record DML and metadata deployment capabilities across all setup scripts. 4. Sanitization: Redacts session tokens in scripts/shared/soap.sh and uses ElementTree for safe metadata construction.
- [SAFE]: No sensitive data exposure or unauthorized network communication was detected. Security-positive patterns were observed, such as the redact_token helper which sanitizes error messages to prevent session ID leakage, and the use of temporary directories with automatic cleanup for metadata staging.
Audit Metadata