Skip to content

Security: kugamon/rowhouse

SECURITY.md

Security Policy

Reporting a vulnerability

Please do not open a public issue. Report it privately, in any of these ways:

  1. GitHub: this repository's Security tab → Report a vulnerability.
  2. Email: support@kugamon.com, with a subject that starts with [SECURITY].
  3. Kugamon support: Submit a Case and say it is a security report.

Please include:

  • what an attacker could do, and to whose data;
  • the steps to reproduce it, as short as you can make them;
  • the Rowhouse version (Setup → Installed Packages), the Salesforce edition, and whether Lightning Web Security is on (Setup → Session Settings).

We will confirm that we have your report, work with you to understand and fix the problem, and agree with you when it is disclosed. We credit reporters in the release notes unless you would rather stay anonymous.

Supported versions

Security fixes go into the latest released version of the managed package. To get them, install or upgrade from that version's link in CHANGELOG.md; an upgrade keeps your grids. The latest version is 1.5, Winter 2027.

How Rowhouse protects data

  • Everything runs as the user. Every read and write runs in user mode, through Salesforce's own object, field and sharing security. Every Apex class is with sharing; there is no system-mode path.
  • One place decides access. Every object and field check goes through a single class, GridSecurity, so a reviewer can read one file to see them all. An automated check fails the build if any other class makes one.
  • The browser is never trusted. The server checks each save, delete, import and action again against the grid's configuration and the user's permissions, whatever the page sends.
  • Nothing leaves the org. No callouts, no outside services, and no third-party code: every line of JavaScript and Apex in the package is Rowhouse's own.
  • Access is granted, not assumed. Users need the Rowhouse Grid User permission set. Building grids on the Grid Admin tab needs Rowhouse Grid Admin as well, and Customize Application.
  • Diagnostics carry no record data. The optional diagnostic lines hold timings and counts only.
  • Exports are safe to open. CSV exports and copied cells neutralize values that a spreadsheet would run as a formula.

How the code is checked

  • Salesforce Code Analyzer runs on every push with the AppExchange and security rules, and the build fails on any finding. A finding that is not a real problem is suppressed as narrowly as possible, with the reason, in code-analyzer.yml. Its path-based field-security check (Graph Engine) runs on each release.
  • The project's hard rules (for example: no global Apex, no callouts, every access check through GridSecurity) are enforced by automated checks on every push; CONTRIBUTING.md lists them.
  • Access tests run as a restricted user, not as an administrator, so a missing object, field or sharing check fails a test.
  • Dependencies are watched by Dependabot. None of them ships in the package: they are the build and test tools.

Scan results

verify Graph Engine

The badges show the latest runs: verify on every push (it includes the rule scan), Graph Engine on the latest release.

Scan What it checks When Findings in 1.4
Salesforce Code Analyzer: AppExchange and security rules The rules the AppExchange security review runs, over the whole package Every push None
Salesforce Code Analyzer: Graph Engine Every path from each server method Rowhouse's components call, for a missing object or field check and for sharing Each release None

Each run's full report is kept with the run on GitHub Actions, under Artifacts.

What is set aside, and why (all of it in code-analyzer.yml):

  • One suppression. ProtectSensitiveData flags the field Grid_User_Layout__c.Key__c because its name looks like a secret's. It holds the owner's user Id and the grid's API name, which make each user's layout of each grid one record. The suppression covers that one field and at most one finding.
  • One rule replaced. PMD's EmptyCatchBlock treats a catch that holds only a comment as empty. Rowhouse's empty catches are deliberate fallbacks, each explained in a comment, so the same rule runs in its place with comments allowed: a catch that is empty and unexplained still fails.
  • No inline suppressions in the Apex code.

There aren't any published security advisories