Skip to content
Back to skills

Ops

ASecurity

Deploy applications to modern cloud platforms (Vercel, Railway, AWS).

  • 23 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 14, 2026
ai-agentsgosqlnextjsrailsdockerawstestingapidatabasefrontend

Works with

  • cli
  • api

Security analysis

A100/100

Pro scans all 21 files and shows the line behind each finding

Scanned October 14, 2026

npx -y skills add andreibesleaga/GABBE --skill ops --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Ops?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Ops
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/andreibesleaga-ops-a1d9acf1/badge)](https://www.skillsdirectory.com/skills/andreibesleaga-ops-a1d9acf1)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: cloud-deploy
description: Deploy applications to modern cloud platforms (Vercel, Railway, AWS).
triggers: [deploy, vercel, railway, aws, sst, serverless, lambda, flightcontrol, neon, supabase]
tags: [ops]
context_cost: medium
---
# Cloud Deploy Skill

## Goal
To configure rapid, developer-friendly deployments on modern PaaS and Serverless platforms.

## Capabilities

### 1. Vercel (Frontend & Edge)
-   **Config**: `vercel.json` for headers, redirects, and clean URLs.
-   **Edge Functions**: Deploying middleware and lightweight backend logic.
-   **Preview Mode**: Setting up preview comments on PRs.

### 2. Railway (Backend & DBs)
-   **Config**: `railway.toml` for build commands and monorepo root.
-   **Databases**: Provisioning Postgres, Redis, MySQL plugins.
-   **Variables**: Managing secrets and environment variables.

### 3. AWS (SST / CDK)
-   **SST**: Defining `sst.config.ts` for serverless stacks (Next.js, API Gateway, DynamoDB).
-   **Constructs**: Using high-level constructs for easy setup (Cron, Bucket, Queue).
-   **IAM**: Setting up fine-grained permissions for functions.

## Steps
1.  **Platform Choice**:
    -   **Frontend/Fullstack JS**: Vercel.
    -   **Docker/Stateful/Backend**: Railway.
    -   **Enterprise/Complex/Serverless**: AWS (via SST).
2.  **Environment Parity**: Ensure `.env` vars match across local and cloud (but values differ).
3.  **Infrastructure as Code**: Always define the deployment config locally (don't click-ops settings).

## Deliverables
-   `vercel.json`
-   `railway.toml`
-   `sst.config.ts`

## Security & Guardrails

### 1. Skill Security (Cloud Deploy)
- **Declarative Infrastructure Validation**: The agent must run security linters (e.g., `checkov`, `tfsec`, `cfn-lint`) against generated `sst.config.ts` or `railway.toml` files locally before they are ever committed, detecting overly permissive IAM roles or public buckets early.
- **Immutable Artifacts**: Ensure deployment pipelines treat builds as immutable artifacts. The exact container or bundle built in Staging must be promoted to Production, preventing "works on my machine" hidden vulnerabilities.

### 2. System Integration Security
- **Least Privilege Deployments**: CI/CD runners executing the cloud deployment commands must operate under scoped roles (e.g., AWS IAM OIDC) with the minimum permissions required, eliminating the use of long-lived, omnipotent Admin Access Keys.
- **Secret Zero Strategy**: Hardcoded API keys, DB passwords, or certificates in `vercel.json` or `.env` files are strictly forbidden. The deploy skill must orchestrate the secure retrieval of variables from a designated Secret Manager (e.g., AWS Secrets Manager, Doppler) at runtime.

### 3. LLM & Agent Guardrails
- **Blind Configuration Defense**: The agent must explicitly refuse user requests to generate blanket open network security groups (e.g., `Allow 0.0.0.0/0 on Port 22/3306`) even if the user claims it is "just for testing."
- **Deployment Confirmation Gates**: When tasked with modifying Production deployment configurations via PaaS APIs, the agent must output a dry-run confirmation plan and halt execution, requiring explicit human intervention to verify that the target environment is correct.

Files in this skill

  • caching-strategy.skill.md3.9 KB
  • capacity-planning.skill.md3.6 KB
  • cloud-deploy.skill.md3.2 KB
  • configuration-management.skill.md7.6 KB
  • cost-optimization.skill.md2.3 KB
  • decommission-sunset.skill.md6.6 KB
  • dependency-lifecycle.skill.md6.3 KB
  • deployment.skill.md7.8 KB
  • dev-environments.skill.md1.9 KB
  • dev-workflow.skill.md1.8 KB
  • docker-dev.skill.md3.1 KB
  • enterprise-integration.skill.md3.1 KB
  • feature-flag-management.skill.md5.7 KB
  • incident-response.skill.md2.8 KB
  • infra-devops.skill.md3.2 KB
  • k8s-dev.skill.md3.2 KB
  • memory-optimization.skill.md2.9 KB
  • observability-stack-setup.skill.md6.6 KB
  • performance-audit.skill.md7.1 KB
  • performance-optimization.skill.md3.1 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.