Before You Deploy: 10 Web App Safety Checks
Ten offline checks to run before you ship a web app. They read your project files, explain every finding with the file and line, and give a safer fix: secrets in code and env files, Dockerfile and GitHub Actions risks, npm package scripts, Content Security Policy, webhook signature checks, Supabase config and functions, .gitignore coverage, and personal data written to logs. Each skill works on its own with Python 3.9+ (standard library, no network). For solo developers and small teams who want one pass over the common mistakes before launch.
عشر أدوات فحص تعمل بلا إنترنت قبل أن تنشر تطبيق الويب. تقرأ ملفات مشروعك، وتشرح كل ملاحظة بالملف والسطر، وتقترح الإصلاح الأسلم: الأسرار المكتوبة في الكود وملفات env، ومخاطر Dockerfile وGitHub Actions، وسكربتات حزم npm، وسياسة أمان المحتوى CSP، والتحقق من توقيع webhooks، وإعدادات Supabase ودوالها، وتغطية ملف .gitignore، والبيانات الشخصية المكتوبة في السجلات. كل أداة تعمل وحدها ببايثون 3.9 أو أحدث. للمطورين المستقلين والفرق الصغيرة الذين يريدون مراجعة واحدة للأخطاء الشائعة قبل الإطلاق.
Works with any agent that signs in to Agensi, such as Claude, ChatGPT, Claude Code, Cursor and Codex CLI.
One payment, lifetime access. 10 skills unlock instantly in your library.
What's included
10 skillsFind API keys, tokens, private keys and database passwords in a project before you commit, publish or zip it: 18 providers by their real key formats, plus whether .gitignore keeps each file with a secret out of git.
Keep .env.example in step with your code and keep .env files out of git: missing and stale variables, real secrets in the example file, unignored .env files and secrets with public prefixes.
Catch Dockerfile mistakes before you build or push: secrets baked into ENV or ARG, curl-pipe-to-shell downloads, copied .env files and keys, root containers and unpinned base images.
Find the GitHub Actions mistakes that let a pull request or issue run code with your secrets: pull_request_target misuse, script injection, hard-coded tokens, unpinned actions and broad permissions.
Check package.json and .npmrc for supply-chain risks: install scripts that download code, http dependencies, leaked registry tokens, wildcard versions, git dependencies and missing lockfiles.
Check Content-Security-Policy headers and meta tags before you ship: script sources open to any host, unquoted keywords and typos, fixed nonces, unsafe-inline, missing object-src, base-uri and frame-ancestors, and parts browsers ignore.
Check webhook endpoints for GitHub, Shopify, Twilio, Paddle, Clerk, Razorpay and 9 more providers: missing signature checks, secrets in code, checks that fail open, == comparisons, HMACs over parsed JSON, replays and duplicate events.
Check Supabase projects before you deploy: service keys in NEXT_PUBLIC_ or VITE_ variables, secrets in config.toml, redirect wildcards that accept any site, open Edge Functions using the service key, and weak auth settings.
Check that .env files, private keys, key stores, tokens and Terraform state stay out of git, Docker images and npm packages, and that node_modules, build output, logs and editor junk are ignored.
Check logging calls and settings before they ship: passwords, tokens, connection strings, Authorization headers and card numbers in logs, login bodies, emails and ID numbers, whole user objects, DEBUG in production, Sentry PII.
See it in action
Hardcoded Secrets Check
Example session with this skill installed
Check this project with hardcoded-secrets-check.
- Read your context and instructions
- Compiled the hardcoded secrets check
$ python3 scripts/check_secrets.py examples/risky
BLOCK .env.production:2 H04 PAYMENT_OR_MESSAGING_TOKEN
Stripe live secret key: STRIPE_SECRET_KEY=<redacted>
Why: A live payment or messaging credential is written in the file: a Stripe live secret or restricted key (sk_live_, rk_live_) or webhook secret (whsec_; Stripe says only pk_ keys are safe to expose), a Slack token (xoxb-, xoxp-, xapp-, xoxe.) or webhook URL, a Discord bot token or webhook URL, a Telegram bot token (bot ID, a colon, then 35 characters starting AA), a Twilio Auth Token (32 characters next to an AC Account SID) or a SendGrid key (SG., 69 characters). Someone can charge cards, issue refunds, read customer data, forge webhooks or message as you.
Safer: Roll the key in the provider's dashboard now (Stripe API keys, Slack app settings, Discord Developer Portal, BotFather /revoke, Twilio Console, SendGrid API Keys) and load the new one from the environment or a secret manager on the server.
BLOCK .env.production:3 H06 PASSWORD_IN_CONNECTION_STRING
BLOCK .npmrc:2 H02 CODE_OR_PACKAGE_TOKEN
BLOCK backend/settings.py:11 H01 CLOUD_ACCESS_KEY
BLOCK deploy/deploy_key.pem:1 H05 PRIVATE_KEY
BLOCK ml/train.py:4 H03 AI_API_KEY
REVIEW .env.production H14 SECRETS_FILE_NOT_IGNORED
REVIEW .npmrc H14 SECRETS_FILE_NOT_IGNORED
REVIEW backend/settings.py:8 H12 SECRET_NAMED_LITERAL
REVIEW backend/settings.py:14 H13 JWT_IN_FILE
REVIEW deploy/deploy_key.pem H14 SECRETS_FILE_NOT_IGNORED
REVIEW server/billing.js:2 H11 TEST_MODE_KEY
REVIEW web/src/firebase.js:4 H15 CLIENT_KEY_NEEDS_RESTRICTION
(Each finding also has Why and Safer lines, as in the first one.)
Verdict: BLOCK. Checked 9 files, 9 with text content.
This is a static check of file text. NOT FLAGGED is not a guarantee.
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 these skills and uses them on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every skill in this bundle goes through our trust checks.
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.