Works with the AI Agents you already use
Web App Security & Bug Auditor (Firebase/Supabase)
Audit web and mobile codebases for security vulnerabilities, leaked secrets, and deployment-breaking bugs.
$7
Web App Security & Bug Auditor (Firebase/Supabase)
Example session with this skill installed
Audit my Next.js project in the current directory for any security risks or deployment-breaking bugs before I push to Vercel.
- Read your context and instructions
- Compiled the web app security
I have completed a security and bug audit of the project. The scan identified one critical credential leak and two high-severity database rule issues.
Security & Bug Audit: nextjs-portfolio-app
Scope: Next.js, Tailwind, Supabase, Vercel | Method: automated scan + manual review | Date: 2024-10-24
Verdict
Do not deploy. A service role key is committed to the frontend bundle, allowing anyone to bypass all database security. Rotate this key immediately.
| Severity | Confirmed | Likely |
|---|---|---|
| Critical | 1 | 0 |
| High | 1 | 1 |
| Medium | 0 | 0 |
| Low | 2 | 0 |
Fix first (in this order)
- Rotate the Supabase
service_rolekey in the dashboard. - Move
SUPABASE_SERVICE_ROLE_KEYto Vercel environment variables and out of.env.local. - Enable RLS on the
profilestable.
Findings
[CRITICAL] Leaked Service Role Key — Confirmed
Where: lib/supabaseClient.js:4
What is wrong: The service_role key is used in a client-side initialization script.
Impact: This key bypasses Row Level Security. Any user can read, edit, or delete the entire database.
Fix:
// lib/supabaseClient.js
import { createClient } from '@supabase/supabase-js'
// Use the public anon key for client-side
export const supabase = createClient(
process.env.NEXT_PUBLIC_SUPABASE_URL,
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY
)
[HIGH] Missing Row Level Security — Confirmed
Where: supabase/migrations/01_init.sql:12
What is wrong: Table profiles is created but RLS is never enabled.
Impact: Even with the correct keys, the table defaults to public access if RLS is not explicitly turned on.
Fix:
ALTER TABLE profiles ENABLE ROW LEVEL SECURITY;
Not verified / needs your check
- Check the Vercel dashboard to ensure
NODE_ENVis set toproductionto prevent source map exposure.
Next steps
- I can help you generate the git command to purge the leaked key from your history.
- Would you like me to apply the RLS fix to your migration file now?
- I can re-run the scan once the environment variables are moved.
Connects securely to your tools. The creator never sees your data.
About this skill
You are one accidental git commit away from a database breach or a broken production build. Most developers only find their security holes after a leak occurs or when a user reports a critical data bug. This skill acts as an automated auditor and manual security partner that finds the vulnerabilities you missed before you deploy.
What it does
- Secrets detection scans for leaked API keys, service roles, and private credentials across your entire codebase.
- Database hardening audits Firebase rules and Supabase RLS policies to prevent unauthorized data access or deletion.
- Bug hunting identifies logic errors like infinite re-renders, async/await mistakes, and memory leaks.
- Mobile audit checks Capacitor and Android configurations for production readiness and security.
- Auth verification tests for IDOR vulnerabilities where users might access or modify data belonging to others.
How it works
- Scope the stack by analyzing
package.json, environment files, and framework-specific configs to identify the attack surface. - Execute scan using the integrated Python scanner to surface 100+ known vulnerability patterns and secret signatures.
- Verify findings by tracing code paths to distinguish between public-by-design keys and critical leaks.
- Manual review against reference checklists for complex authorization logic that static tools cannot see.
- Generate report providing a severity-ranked list of confirmed issues with copy-paste code fixes.
Frameworks & tools
Built for React, Next.js, Node/Express, and Cloudflare Workers. Specialized support for Firebase (Firestore/Storage/Realtime DB), Supabase (RLS/SQL), and mobile projects using Capacitor or Android.
Why this beats prompting it yourself
Generic prompts often return "security theater" like warnings about public keys that are meant to be public. This skill uses a specialized scanner and stack-specific reference checklists to filter noise, focusing only on verified, exploitable paths with high-fidelity fixes.
Use cases
- Audit a project before a hackathon submission or client handover.
- Check if your Supabase or Firebase rules are actually protecting user data.
- Find why an app works on a local machine but crashes in the production APK.
- Clean up leaked credentials and harden environment variable hygiene.
Known limitations
Does not perform live penetration testing or dynamic runtime analysis. Cannot scan third-party dependencies for CVEs or replace a professional audit for regulated payment or health data.
Static review only; no live probing or runtime testing. Does not scan dependencies for CVEs. Requires Python for automated pass.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 3 days ago
Requires