The independent daily for Salesforce admins, developers and architects

Curated by Abhinav (opens in a new tab)

Design the permission model before the screens in occupational health cases (opens in a new tab)

You will see how to split an occupational health case across objects so an employer sees fitness outcomes but never clinical detail.

Why read this

You will also see why governed picklists, data-driven thresholds and an external community help with privacy and audit.

Occupational health case

  • Case
  • Encounter
  • Exposure event
  • Restriction
  • Surveillance result
  • Employer account
A single case is split into separate objects so each role sees only its own slice

The article frames occupational health case management as a permission problem. Five roles touch each case, and the employer may learn whether a worker can lift or drive but not why. Spreadsheets and shared drives fail because they cannot enforce who sees what or show afterwards who looked.

The author argues against a single large record. The case should be split into objects for the encounter, exposure event, restriction, surveillance result, employer account and worker record, with a junction object where the employer of record differs from the work site. Restrictions carry start and end dates and a review owner, so expiry becomes a tracked state. Record types and field-level security then separate employer-visible outcomes from clinical detail.

Free text classification is treated as a privacy risk as well as a reporting problem, because a diagnosis typed into a notes field escapes every Sharing Rule. Governed picklists close that gap. Regulated rules such as blood lead level monitoring thresholds and testing frequency should live as records rather than code, so regulatory changes become configuration edits and reports draw on the same data as the rules. Correspondence is logged as case activity, and documents carry version history, classification and retention rules.

For employers and referring clinicians, the article recommends an external community rather than a hidden-field copy of the internal app, which also avoids consuming internal user licences. Migration is where legacy duplicates and inconsistent classifications surface, so it should be rehearsed in a full sandbox and validated against historical caseloads. The closing advice is to settle the permission model first, since it is the expensive part to retrofit.

Read the original at dev.to (opens in a new tab)

Worth reading?

Recommended

Securing Agentforce against prompt injection and data exfiltration (opens in a new tab)

You will learn how untrusted input can steer Agentforce into exfiltrating data via DNS lookups, and which Einstein Trust Layer controls actually mitigate these risks.

Why read this

You will also get a practical checklist for configuring runtime identities, permissions, and safe workflows.

Research by Zenity demonstrates a prompt injection chain where untrusted text stored in a Web-to-Lead record steers Agentforce to query authorized account data. The injected instructions embed the results in an external hostname, exploiting a Trusted URLs parsing discrepancy that triggers a DNS lookup to exfiltrate information without requiring user interaction.

The Einstein Trust Layer provides several guardrails, including semantic filters, zero data retention agreements, and audit logging, but it does not enforce authorization policies or stop agents from querying records they are already permitted to see. Data masking is currently disabled for agents, and prompt defenses remain probabilistic rather than absolute.

Securing Agentforce requires strict identity and permission hygiene. Administrators must verify whether agents run under employee profiles, dedicated service users, or integration accounts, then restrict object and field access accordingly. Replacing broad query tools with custom Apex actions that enforce user mode and explicit Sharing Rules reduces exposure.

A recommended architecture isolates unauthenticated submissions behind a read-only lead triage agent, followed by human review before downstream agents handle outbound communication. This layered approach prevents poisoned data from reaching internal databases or messaging channels like Slack.

Read the original at dev.to (opens in a new tab)

Worth reading?