How do I install the Tenant Scope Audit skill?

Run npx skillmds@latest add amineorion/tenant-scope-audit in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Tenant Scope Audit safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Tenant Scope Audit?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Tenant Scope Audit free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Tenant Scope Audit?

amineorion (@amineorion) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Tenant Scope Audit by amineorion — Agent Skill

Name: Tenant Scope Audit
Author: amineorion

Tenant Scope Audit

Audit a multi-tenant codebase for queries, caches, jobs, or searches that are missing a tenant filter — the missed WHERE tenant_id that becomes a cross-tenant data leak. Use whenever the user works on a multi-tenant SaaS, asks to check tenant isolation, reviews data-access code, or mentions tenant scoping, RLS, or data leaks.

amineorion 75f24f2 2 files · 3.8 KB Updated Sep 17, 2026

File contents

Tenant-scope audit

Find data access that is not scoped to a tenant. In a shared-schema multi-tenant system, one unscoped query is a data breach, not a performance bug.

Run the scan

bash ${CLAUDE_SKILL_DIR}/scripts/scan-tenant-scope.sh

Scope it to a path or the current diff when reviewing a change:

bash ${CLAUDE_SKILL_DIR}/scripts/scan-tenant-scope.sh api/ workers/

Set TENANT_COL if the tenant column isn't tenant_id|org_id|account_id|workspace_id.

Then reason about each hit

The scanner is a heuristic. For every flagged line, decide which case it is:

Genuinely unscoped (fix it) — a query that returns rows across tenants. The fix is a tenant filter, or routing through the tenant-safe data layer.
Scoped indirectly (verify, then note) — row-level security, a for_tenant() repository, or a current_setting('app.tenant_id') is enforcing it. Confirm the enforcement actually covers this path.
Intentionally global (document it) — an admin/reporting query that must span tenants. It needs an explicit comment and ideally a separate, audited code path.

Look past the database — RLS does not cover these

Row-level security is a safety net, not a fortress wall. Check the vectors it misses:

Caches — is the cache key namespaced by tenant? A shared key serves tenant A's data to tenant B.
Background jobs — are queues/global workers filtering by tenant, or exporting every tenant's rows?
Search indexes — are documents written and queried with a tenant key?
File paths / blob storage — are object keys prefixed by tenant?

Report

Summarize as: confirmed leaks (with file:line and the fix), needs-verification, and documented-global. Lead with anything that returns other tenants' data today.

---name: tenant-scope-auditdescription: Audit a multi-tenant codebase for queries, caches, jobs, or searches that are missing a tenant filter — the missed WHERE tenant_id that becomes a cross-tenant data leak. Use whenever the user works on a multi-tenant SaaS, asks to check tenant isolation, reviews data-access code, or mentions tenant scoping, RLS, or data leaks.---# Tenant-scope auditFind data access that is not scoped to a tenant. In a shared-schema multi-tenantsystem, one unscoped query is a data breach, not a performance bug.## Run the scan```bashbash ${CLAUDE_SKILL_DIR}/scripts/scan-tenant-scope.sh```Scope it to a path or the current diff when reviewing a change:```bashbash ${CLAUDE_SKILL_DIR}/scripts/scan-tenant-scope.sh api/ workers/```Set `TENANT_COL` if the tenant column isn't `tenant_id|org_id|account_id|workspace_id`.## Then reason about each hitThe scanner is a heuristic. For every flagged line, decide which case it is:1. **Genuinely unscoped (fix it)** — a query that returns rows across tenants. The fix is a tenant filter, or routing through the tenant-safe data layer.2. **Scoped indirectly (verify, then note)** — row-level security, a `for_tenant()` repository, or a `current_setting('app.tenant_id')` is enforcing it. Confirm the enforcement actually covers this path.3. **Intentionally global (document it)** — an admin/reporting query that must span tenants. It needs an explicit comment and ideally a separate, audited code path.## Look past the database — RLS does not cover theseRow-level security is a safety net, not a fortress wall. Check the vectors it misses:- **Caches** — is the cache key namespaced by tenant? A shared key serves tenant A's data to tenant B.- **Background jobs** — are queues/global workers filtering by tenant, or exporting every tenant's rows?- **Search indexes** — are documents written and queried with a tenant key?- **File paths / blob storage** — are object keys prefixed by tenant?## ReportSummarize as: confirmed leaks (with file:line and the fix), needs-verification, anddocumented-global. Lead with anything that returns other tenants' data today.

Frequently asked questions