agentleFS
toby7979toby7979/sealmyapp-tools/skills/vibe-security/SKILL.md

vibe-security

Skill0 starsChanged 32 days agoMITView on GitHub

Security guardrails for AI-built (vibe-coded) web apps using Supabase, Firebase, Stripe or AI APIs. Use when writing or reviewing auth, database queries, Supabase RLS policies, Firebase rules, migrations, environment variables, API keys or payment code, or before launching an app built with Lovable, Bolt, Cursor, Replit or v0.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. raroque/vibe-security-skill★ 996same name
    2. Falcon305/vibe-starter★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repositoryOne other file it gives its agents.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 3 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Never put a secret in client code or in a browser-exposed variable (NEXT_PUBLIC_, VITE_, EXPO_PUBLIC_, PUBLIC_, REACT_APP_).

    line 15
  2. Distinctive:

    Call paid APIs from server code that checks the signed-in user and rate-limits.

    line 18
  3. Distinctive:

    When checking whether a key is safe, look only at its prefix.

    line 20
  4. Very specific:

    Use owner-based policies with (select auth.uid()), scoped to authenticated.

    line 24
  5. Very specific:

    Never fix empty results or permission errors with using (true), with check (true), disabling RLS or granting to anon.

    line 25
  6. Very specific:

    Keep the service_role / secret key server-only.

    line 27
  7. Very specific:

    No test-mode rules (allow read, write: if true.

    line 43
  8. Distinctive:

    Every API route, server action and edge function checks the signed-in user and ownership.

    line 47

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Vibe-coded app security

Apply these rules whenever you touch keys, env vars, database access, auth or payments. If the user asks for something that breaks them, explain the risk and offer the safe alternative.

1. Secrets stay on the server

  • Never put a secret in client code or in a browser-exposed variable (NEXT_PUBLIC_, VITE_, EXPO_PUBLIC_, PUBLIC_, REACT_APP_).
  • Secret: Stripe sk_ / rk_ / whsec_, OpenAI or Anthropic sk- / sk-ant-, Supabase service_role / sb_secret_, AWS AKIA, GitHub ghp_ / github_pat_, Resend re_, SendGrid SG., database URLs with passwords, private keys.
  • Public by design: Supabase anon / sb_publishable_ key, Stripe pk_ key, Firebase web config apiKey (but a Gemini AIza key is secret).
  • Call paid APIs from server code that checks the signed-in user and rate-limits.
  • A secret that was ever in the frontend or a public repo must be rotated first (new key, revoke old). Removing it from the code is not enough.
  • When checking whether a key is safe, look only at its prefix; don't ask the user to paste a whole secret.

2. Supabase: RLS on, owner policies, no using (true)

  • Enable RLS on every public table.
  • Use owner-based policies with (select auth.uid()), scoped to authenticated; update policies need both using and with check.
  • Never fix empty results or permission errors with using (true), with check (true), disabling RLS or granting to anon. Find out why the owner check fails (not signed in, wrong column) and fix that.
  • using (true) is only OK on SELECT for genuinely public data, never for writes or personal data.
  • Keep the service_role / secret key server-only.
  • Users must not be able to update their own role, plan, credits or admin flags.
  • API-exposed views use security_invoker = true; private files go in private buckets.

Example owner policy:

alter table public.notes enable row level security;
create policy "notes: read own" on public.notes
  for select to authenticated using ((select auth.uid()) = user_id);
create policy "notes: update own" on public.notes
  for update to authenticated
  using ((select auth.uid()) = user_id) with check ((select auth.uid()) = user_id);

3. Firebase: lock the rules

  • No test-mode rules (allow read, write: if true; or the date-limited test rule).
  • Rules check request.auth != null and request.auth.uid against the owner field.

4. Server code and payments

  • Every API route, server action and edge function checks the signed-in user and ownership; admin routes check the role on the server.
  • Prices and paid status come from the server or a verified payment webhook, never from the client.

Free tools

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.