The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides several automation scripts (scripts/configure-named-credential.sh, scripts/set-api-credential.sh) designed to help users configure Salesforce credentials via the Salesforce CLI (sf). These scripts are intended for local execution by the user to avoid manual UI steps.
[DYNAMIC_EXECUTION]: The script configure-named-credential.sh dynamically generates Apex code at runtime and executes it using the sf apex run command. This is used to programmatically store encrypted API keys within Salesforce's Enhanced Named Credentials architecture. This behavior is the primary intended purpose of the automation utility and is performed locally within the user's authenticated environment.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill handles sensitive integration data, such as API keys and OAuth parameters. It adheres to secure practices by using read -s to hide user input in the terminal and by leveraging Salesforce's platform-level encryption (AES-256) for storing secrets in Named Credentials. The skill documentation explicitly warns against hardcoding credentials and provides templates for secure credential management.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as OpenAPI specifications and user-provided integration requirements. To mitigate risks, it includes a proactive security validator (hooks/scripts/validate_integration.py) that scans all generated Apex code for hardcoded secrets, missing try-catch blocks, and unsafe DML/SOQL patterns in loops before deployment.
[SAFE]: No malicious patterns, prompt injections, or unauthorized data exfiltration attempts were detected. The skill's architecture prioritizes the principle of least privilege (e.g., minimizing OAuth scopes) and secure secret management.

sf-integration