Open beta — every feature free with your own AI API key. Beta software: expect (and report!) rough edges.

Security Review

A Salesforce security audit you can run this afternoon

Fourteen modules sweep your org — privileged access, sharing model, field-level security, guest access, Apex, integrations and more — combining deterministic checks with AI analysis to produce a scored report, triaged findings with evidence, and a manual-review guide for the judgement calls.

What you can do

  • Fourteen audit modules: org baseline, authentication & SSO, users & licences, privileged access, permissions hygiene, object & field-level security, sharing model, guest & external access, Apex security, automation & flows, integrations, packages & platform, files & data egress, monitoring & audit
  • Deterministic checks find the facts; AI analysis groups the evidence and explains the risk
  • Every finding carries severity, evidence, and a link to the metadata it came from
  • A coverage matrix shows exactly which modules ran and what each one examined
  • Triage findings as resolved, accepted, or false-positive — state persists for your whole team
  • Export Markdown or CSV for stakeholders, plus an executive summary
  • A manual-review guide for the calls a tool should not make alone
  • Runs as a detached background job — start it and leave the page

How it works

1

Start a review

One click on the org’s Security Review page. The run detaches, so you can close the tab and come back to it.

2

Watch coverage fill in

The matrix ticks off modules as they complete, and findings stream in as they are discovered.

3

Triage with evidence

Findings arrive severity-ranked with the underlying metadata attached. Resolve, accept, or dismiss — your team sees the same state.

4

Export and act

Markdown or CSV for the stakeholders who asked, and the manual-review guide for the human judgement calls.

Why it's different

Evidence, not a checklist score

Salesforce Health Check gives you a percentage against a baseline. This walks fourteen modules of real metadata, has AI reason over the collected evidence in groups, and hands you findings you can act on — each with the record, field, or profile behind it. The audit itself only ever reads: it inspects your settings, it never changes them.

Try asking

  • Audit an inherited org before you agree to own it
  • Answer “who can see what?” with evidence instead of guesses
  • Produce a security report for a client, an auditor, or a board pack
  • Catch over-permissioned profiles and stale integration users
  • Re-run quarterly and compare what changed

Security Review — questions

Put Security Review to work on your org

Start free with the Metadata Explorer — every AI feature is free while the beta runs, and a competitive price is still being worked out.