Made several minor refinements to improve clarity and specificity while preserving all original points: (1) Changed 'supported APIs' to 'platform-supported APIs' for clarity about Salesforce platform capabilities, (2) Added 'and use sharing keywords appropriately' to the authentication section to acknowledge proper sharing enforcement, (3) Added 'requirements' to clarify what should be documented for customers, (4) Added 'and third-party providers to customers' to be more specific about disclosure obligations, (5) Expanded 'code injection or CSRF attacks' to 'code injection, SOQL injection, XSS, and CSRF attacks' to be more comprehensive about common attack vectors, (6) Added 'with Named Credentials where possible' to OAuth guidance for better security practice, (7) Added 'to avoid known vulnerabilities' to dependency management to clarify the security purpose. All security rules were selected based on their direct relationship to the architectural principles discussed: ApexBadCrypto relates to the encryption/cryptography principle, ApexCRUDViolation and ApexSharingViolations relate to authentication/authorization and least privilege principles, ApexCSRF, ApexSOQLInjection, ApexXSSFromEscapeFalse, ApexXSSFromURLParam, and VfCsrf/VfUnescapeEl all relate to the secure coding practices principle covering various attack vectors, ApexInsecureEndpoint relates to endpoint security, ApexSuggestUsingNamedCred relates to the third-party integrations principle, hardcoded credentials rules relate to sensitive data handling, remote site security rules relate to third-party integrations, and vulnerability detection rules relate to the dependency management principle.