sanaskills

Sixty ways to make an AI coding assistant think like a senior engineer.

Written skill files for code review, debugging, testing, API and database design, frontend, and DevOps — each one carries diagnostic heuristics and named anti-patterns, not just a task description. Paired with a small, read-only MCP server that gives an assistant safe local project context.

Get the catalog — $9.99 Browse the 60 skills

One-time fee of $9.99. Lifetime access. Instant download. No subscription.

skills/34-api-security-auditor/SKILL.md 34 / 60

API Security Auditor

Review an API design for common security weaknesses.

Checks — auth, object-level access, mass-assignment, rate limits

Red flag

  • Ownership check missing on a resource-by-id endpoint
  • Password reset endpoint with no rate limiting
  • Stack traces exposed in client-facing error responses

the catalog

Sixty skills, grouped by the work you actually do.

Every file follows the same shape — purpose, workflow, output contract, diagnostic heuristics, red flags, a worked example — applied to a different corner of the job.

Code & correctness 10
  • Code Review Pro
  • Bug Hunter
  • Refactoring Strategist
  • Clean Code Auditor
  • Legacy Modernizer
  • Code Simplifier
  • Performance Optimizer
  • Error Diagnosis
  • Exception Design
  • Type Safety Auditor
Git & collaboration 10
  • Git Commit Writer
  • Pull Request Reviewer
  • PR Description Writer
  • Issue Triage
  • Git Conflict Helper
  • Changelog Generator
  • Release Notes Builder
  • Branch Strategy Advisor
  • Repository Auditor
  • Issue-to-Code Planner
Testing & QA 10
  • Unit Test Builder
  • Integration Test Builder
  • E2E Test Designer
  • Test Coverage Analyzer
  • Test Case Designer
  • Regression Planner
  • API Test Generator
  • Bug Reproduction Assistant
  • QA Checklist Generator
  • Test Failure Analyzer
APIs & data 10
  • REST API Designer
  • API Documentation Writer
  • OpenAPI Contract Builder
  • API Security Auditor
  • Backend Architecture Reviewer
  • Database Schema Designer
  • SQL Query Optimizer
  • Database Migration Planner
  • Database Debugger
  • Cache Strategy Advisor
Frontend 10
  • React Architecture Expert
  • Next.js Builder
  • UI Component Generator
  • Accessibility Auditor
  • Responsive UI Reviewer
  • Frontend Performance Optimizer
  • Tailwind UI Specialist
  • State Management Advisor
  • Form Validation Builder
  • Frontend Bug Hunter
DevOps & release 10
  • Dockerfile Optimizer
  • Docker Compose Builder
  • CI/CD Pipeline Builder
  • GitHub Actions Expert
  • Kubernetes Troubleshooter
  • Terraform Reviewer
  • Deployment Checklist
  • Environment Config Auditor
  • Observability Designer
  • Production Readiness Auditor

the toolkit

A small MCP server that only ever reads.

Nine tools giving an assistant local project context — no shell execution, no file writes, nothing that touches your system beyond reading what you point it at.

list_files / read_text_file / search_text

Basic project navigation and text search across a directory tree.

project_overview

Detects common project markers — package.json, Dockerfile, CI config — and summarizes the top level.

git_diff_stat

Returns a diff summary for the working tree of a git repository.

find_todos

Collects TODO / FIXME / HACK / XXX comments into one list.

scan_for_hardcoded_secrets

Pattern-based check for likely API keys, tokens, and private key blocks — a first pass, not a replacement for a dedicated scanner.

dependency_manifest_summary

Counts declared dependencies across npm, pip, pyproject, Go, and Cargo manifests.

code_size_report

Largest files by line count, grouped by extension — a rough map of where a codebase's weight sits.

pricing

One price. Everything above.

$9.99

All 60 skill files, the MCP server, setup notes, and updates published to this version line.

One-time feeLifetime accessInstant download
Pay once, keep it forever.

before you buy

This pack is built around 60 written skill files — focused instruction sets that teach your AI coding assistant to reason like a senior engineer on a specific job: reviewing a pull request, hunting a bug, designing an API, hardening a Dockerfile. Each one is genuinely useful on its own, and paired with your assistant it can save real hours across code review, debugging, testing, and DevOps.

In plain terms: this is a set of Markdown skill files and a small Python MCP server — not a hosted app or an official Anthropic, OpenAI, GitHub, or Model Context Protocol product. MCP refers to the open Model Context Protocol standard, which this toolkit implements a server for. What you get out of it still depends on the AI assistant you pair it with — the skills guide the reasoning, they don't replace the assistant, and there's no guarantee of a specific coding outcome.