The Agent Skills Directory

[SAFE]: The skill provides standard documentation and code snippets for interacting with Salesforce via the jsforce and simple-salesforce libraries, which are widely recognized and reputable.
[SAFE]: It correctly implements credential management by utilizing environment variables (SF_USERNAME, SF_PASSWORD, SF_SECURITY_TOKEN) to avoid the risk of hardcoded secrets.
[SAFE]: All network activity is directed toward official Salesforce domains (login.salesforce.com) and API services, consistent with the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from Salesforce records (e.g., Account names), which is an inherent ingestion surface for this type of tool.
Ingestion points: Data is fetched via SOQL queries and record retrieval in SKILL.md.
Boundary markers: None are present in the code snippets, which is typical for a basic 'hello world' example.
Capability inventory: The skill has Read, Write, and Edit tool permissions for record management.
Sanitization: The example focuses on API usage and does not include data sanitization logic, relying on standard library handling.

salesforce-hello-world