Quick Start

Get a Manicode code security prompt running in your workflow in minutes.

Step 1: Find Your Prompt

Browse the Prompt Library (328 code security prompts) or the Skills Library (267 skills with structured metadata). Each targets a specific framework.

If you work with...	Look for...
Node.js / Express / Fastify	Node.js code security prompt or skill
Python / Django / Flask / FastAPI	Python code security prompt or skill
Java / Spring Boot	Java code security prompt or skill
Go	Go code security prompt or skill
React / Next.js	React or Next.js code security prompt or skill
.NET / ASP.NET Core	.NET code security prompt or skill
Ruby on Rails	Ruby on Rails code security prompt or skill
Rust	Rust code security prompt or skill

Step 2: Choose Your Deployment Method

Pick the method that matches how you work.

Option A: Paste into a Web Chat Interface

The simplest approach. Open your preferred AI chat interface and paste the prompt into the system instructions.

ChatGPT:

1. Create a new Project in ChatGPT
2. Paste the prompt into the Project instructions
3. All conversations in that project will enforce the security rules

Claude:

1. Create a new Project on claude.ai
2. Paste the prompt into the project instructions
3. All conversations in that project use the security prompt

Google AI Studio:

1. Create a new prompt in AI Studio
2. Paste the prompt into the System Instructions field

See Web Chat Interfaces for detailed setup instructions.

Option B: Add to Your IDE

Most AI-powered IDEs support project-level custom instructions. Create the appropriate config file in your repository root:

IDE / Tool	File to create
GitHub Copilot	.github/copilot-instructions.md
Cursor	.cursor/rules/security.mdc
Windsurf	.windsurfrules
JetBrains AI	.junie/guidelines.md
Cline / Roo Code	.clinerules
Amazon Q Developer	.amazonq/rules/security.md

Paste the full prompt or SKILL.md content into the file and commit it. Every developer on the team gets the security rules automatically.

See IDE Integrations for detailed setup per tool.

Option C: Configure Your Coding Agent

AI coding agents that operate autonomously benefit the most from security prompts, since they make decisions and write code without constant human oversight.

Agent	File to create
Claude Code	CLAUDE.md
OpenAI Codex	AGENTS.md
Aider	CONVENTIONS.md
Windsurf Cascade	.windsurfrules
Cursor Agent	.cursor/rules/security.mdc
Amazon Q CLI	.amazonq/rules/security.md

Paste the full prompt or SKILL.md content into the file and commit it.

See AI Coding Agents for detailed setup per agent.

Option D: Call via API

For automated workflows, CI/CD pipelines, or custom tooling, pass the prompt as the system message in an API call.

OpenAI:

from openai import OpenAI

client = OpenAI()

with open("prompts/node-security.md") as f:
    security_prompt = f.read()

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": security_prompt},
        {"role": "user", "content": "Write an Express.js login endpoint"}
    ]
)

Anthropic:

import anthropic

client = anthropic.Anthropic()

with open("prompts/python-security.md") as f:
    security_prompt = f.read()

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=4096,
    system=security_prompt,
    messages=[
        {"role": "user", "content": "Write a Django authentication view"}
    ]
)

Google Gemini:

from google import genai

client = genai.Client()

with open("prompts/java-security.md") as f:
    security_prompt = f.read()

response = client.models.generate_content(
    model="gemini-2.0-flash",
    config=genai.types.GenerateContentConfig(
        system_instruction=security_prompt
    ),
    contents="Write a Spring Boot REST controller with authentication"
)

See API and Programmatic Usage for more examples including OpenRouter, dynamic prompt selection, and CI/CD integration.

Step 3: Verify It Works

Test the prompt by asking the AI to generate code that typically has security implications:

• "Write a login endpoint" — should include password hashing, rate limiting, and secure session handling
• "Write a file upload handler" — should include file type validation, size limits, and path traversal prevention
• "Write a database query" — should use parameterized queries, not string concatenation

If the generated code includes security controls unprompted, the prompt is working.

Next Steps

• Prompt Library — Browse all 328 code security prompts
• Skills Library — 267 skills with structured metadata and router discovery
• Team and Enterprise Deployment — Scale prompts across your organization
• Prompt Delivery — How Manicode delivers updates to your repo

Limitations

• Prompts influence but do not guarantee model behavior. LLMs are non-deterministic — the same prompt and input can produce different outputs across runs.
• Security prompts reduce but do not eliminate insecure code generation. Always review LLM output before using it in production.
• Prompts do not replace static analysis (SAST), dynamic testing (DAST), dependency scanning, or manual code review.