Back to Skills

security-prompts-controls

harperaa
Updated Yesterday
30 views
0
View on GitHub
Metaaiapidesign

About

This skill provides pre-built security prompt templates for common development tasks like contact forms, user updates, and public APIs. It enables quick implementations by leveraging existing Secure Vibe Coding OS utilities for straightforward features. Use these templates for rapid, standardized security patterns rather than complex scenarios requiring custom prompt engineering.

Documentation

Built-In Controls Security Templates

Purpose

This skill provides simple, fast security prompt templates for common features that leverage existing Secure Vibe Coding OS utilities. These are streamlined templates for straightforward implementations that don't require extensive customization.

When to use these templates:

  • Simple, common features
  • Quick implementations (15-30 minutes)
  • Leveraging existing security utilities
  • Standard patterns without complex requirements

When to use prompt-engineering instead:

  • Complex features requiring multiple security layers
  • Custom security requirements
  • Features needing extensive customization
  • New patterns not covered by utilities

Available Templates

01: Contact Form

File: 01_contact_form.md When to use: Public form with full security stack Time: 15-30 minutes

Security Controls (Pre-configured):

  • CSRF protection (withCsrf)
  • Rate limiting (withRateLimit - 5 per 15 min)
  • Input validation (Zod schemas)
  • XSS sanitization (safeTextSchema)
  • Secure error handling (handleApiError)

Implementation:

  • Single API route
  • Existing middleware composition
  • Standard validation schemas
  • Minimal customization needed

Trigger keywords: "contact form", "simple form", "feedback form", "quick form", "form with security"

Use case: Contact forms, feedback forms, newsletter signups, support requests

Difference from prompt-engineering/01_secure_form.md:

  • Simpler template
  • Less customization guidance
  • Faster implementation
  • Assumes standard use case

02: Authenticated Update

File: 02_authenticated_update.md When to use: User data modification with auth Time: 30 minutes

Security Controls (Pre-configured):

  • Clerk authentication (auth())
  • Rate limiting (withRateLimit)
  • Input validation (Zod schemas)
  • XSS sanitization (safeTextSchema)
  • Secure error handling (handleApiError)

Implementation:

  • Single authenticated API route
  • Standard data update pattern
  • Existing utilities
  • Basic ownership check

Trigger keywords: "authenticated update", "update profile", "user update", "edit profile", "user data modification"

Use case: Profile updates, user settings, preference changes

Important: For resources requiring strict ownership verification, follow up with: → .claude/skills/security/security-prompts/auth-authorization/03_ownership.md

Difference from prompt-engineering/02_authenticated_endpoint.md:

  • Simpler ownership check
  • Less comprehensive authorization
  • Faster for basic updates
  • Standard user → own data pattern

03: Public API

File: 03_public_api.md When to use: Public GET endpoints with validation Time: 20 minutes

Security Controls (Pre-configured):

  • Rate limiting (withRateLimit)
  • Input validation (Zod query schemas)
  • Error handling (handleApiError)
  • Pagination (standard pattern)

Implementation:

  • Read-only endpoints
  • Query parameter validation
  • Standard pagination
  • No authentication required

Trigger keywords: "public API", "public endpoint", "GET endpoint", "read-only API", "public data"

Use case: Blog posts, product catalogs, public listings, documentation

Difference from prompt-engineering/03_public_endpoint.md:

  • Simpler pagination
  • Standard patterns only
  • Less customization
  • Faster for typical GET endpoints

Usage Pattern

1. Identify Template Match

Simple contact form: → 01_contact_form.md ✅

Profile editing: → 02_authenticated_update.md ✅

Public data listing: → 03_public_api.md ✅

Admin features: → Use prompt-engineering/04_admin_action.md ❌ (not here)

File uploads: → Use prompt-engineering/05_file_upload.md ❌ (not here)

2. Quick Implementation

For: [Simple feature name]

Template: built-in-controls/[template].md

Estimated time: [15-30 minutes]

This uses existing Secure Vibe Coding OS utilities:
- [List utilities from template]

Let me implement this quickly...

3. When to Upgrade

Start with built-in-controls, upgrade if needed:

Started with: built-in-controls/02_authenticated_update.md
Need to add: Ownership verification for team resources
Upgrade to: auth-authorization/03_ownership.md
Started with: built-in-controls/01_contact_form.md
Need to add: File attachment support
Upgrade to: prompt-engineering/05_file_upload.md

Template Comparison

Contact Form Templates

built-in-controls/01_contact_form.md:

  • ✅ 15-30 minutes
  • ✅ Standard fields (name, email, message)
  • ✅ Pre-configured controls
  • ❌ Limited customization guidance

prompt-engineering/01_secure_form.md:

  • ⏱️ 30 minutes
  • ✅ Any form fields
  • ✅ Comprehensive security guidance
  • ✅ Extensive customization tips
  • ✅ More testing guidance

Choose built-in-controls when: Standard contact form, quick implementation Choose prompt-engineering when: Custom fields, complex validation, learning security patterns

Authenticated Update Templates

built-in-controls/02_authenticated_update.md:

  • ✅ 30 minutes
  • ✅ User → own data
  • ✅ Basic ownership check
  • ❌ Simple scenarios only

prompt-engineering/02_authenticated_endpoint.md:

  • ⏱️ 45 minutes
  • ✅ Any authenticated operation
  • ✅ Complex authorization
  • ✅ Team/organization resources
  • ✅ Comprehensive audit logging

Choose built-in-controls when: User editing their own profile/settings Choose prompt-engineering when: Team resources, complex permissions, admin features

Public API Templates

built-in-controls/03_public_api.md:

  • ✅ 20 minutes
  • ✅ Simple GET endpoints
  • ✅ Standard pagination
  • ❌ Basic use cases only

prompt-engineering/03_public_endpoint.md:

  • ⏱️ 30 minutes
  • ✅ Complex queries
  • ✅ Advanced pagination
  • ✅ Search/filtering
  • ✅ Performance optimization

Choose built-in-controls when: Simple data listing Choose prompt-engineering when: Search, complex filtering, advanced pagination

Workflow Integration

Quick Feature Flow

1. Quick Implementation
   → built-in-controls/[template].md

2. Test
   → Follow template testing checklist

3. Deploy
   → Update threat model (threat-modeling/08_update_model.md)

Feature Growth Flow

1. Start Simple
   → built-in-controls/[template].md

2. Add Complexity (if needed)
   → Upgrade to prompt-engineering/[template].md
   → Or add auth-authorization/[template].md

3. Test Thoroughly
   → prompt-engineering/08_security_testing.md

4. Review
   → threat-modeling/04_code_review.md

Utilities Reference

All templates use these Secure Vibe Coding OS utilities:

Middleware

withCsrf():

import { withCsrf } from "@/lib/security/csrf";
export const POST = withCsrf(handler);

withRateLimit():

import { withRateLimit } from "@/lib/security/rate-limit";
export const POST = withRateLimit(withCsrf(handler), {
  requests: 5,
  window: 15 * 60 // 15 minutes
});

Validation

safeTextSchema:

import { safeTextSchema } from "@/lib/validation";
const schema = z.object({
  message: safeTextSchema.max(1000)
});

emailSchema:

import { emailSchema } from "@/lib/validation";
const schema = z.object({
  email: emailSchema
});

validateRequest():

import { validateRequest } from "@/lib/validation";
const data = await validateRequest(req, schema);

Error Handling

handleApiError():

import { handleApiError } from "@/lib/errors";
try {
  // Logic
} catch (error) {
  return handleApiError(error);
}

Authentication

auth():

import { auth } from "@clerk/nextjs/server";
const { userId } = auth();
if (!userId) return Response.json({error: "Unauthorized"}, {status: 401});

Testing Checklists

Contact Form Testing

- [ ] Form requires CSRF token
- [ ] Rate limiting blocks 6th submission
- [ ] XSS in message field is sanitized
- [ ] Invalid email rejected
- [ ] Empty fields rejected
- [ ] Form submission succeeds with valid data

Authenticated Update Testing

- [ ] Unauthenticated request rejected (401)
- [ ] User can update own data
- [ ] User cannot update other user's data (403)
- [ ] Invalid data rejected
- [ ] Rate limiting works
- [ ] Success response correct

Public API Testing

- [ ] Endpoint returns data without auth
- [ ] Pagination works correctly
- [ ] Invalid query params rejected
- [ ] Rate limiting enforces limits
- [ ] Error responses secure
- [ ] Performance acceptable

When NOT to Use Built-In Controls

Admin features → Use: prompt-engineering/04_admin_action.md

File uploads → Use: prompt-engineering/05_file_upload.md

Complex authorization → Use: auth-authorization/[appropriate_template].md

Multiple middleware layers → Use: prompt-engineering/06_composable_middleware.md

Custom security patterns → Use: prompt-engineering/07_new_control.md

Team/organization resources → Use: prompt-engineering/02_authenticated_endpoint.md + auth-authorization/03_ownership.md

Agent Usage

Quick Feature Request

User: "Add a contact form"

Agent Response:
"I'll use the quick contact form template from built-in-controls.
This leverages existing security utilities for a 15-30 minute implementation.

Loading: .claude/skills/security/security-prompts/built-in-controls/01_contact_form.md"

Upgrade Scenario

User: "The contact form needs file attachment support"

Agent Response:
"File uploads require additional security controls beyond the simple template.
I'll upgrade to the comprehensive file upload template.

Switching to: .claude/skills/security/security-prompts/prompt-engineering/05_file_upload.md"

Best Practices

Start Simple

Begin with built-in-controls:

  • Faster implementation
  • Less complexity
  • Standard patterns
  • Easy to understand

Upgrade when needed:

  • Requirements grow
  • Need custom patterns
  • Complex authorization
  • Multiple security layers

Always Test

Even simple templates need testing:

  1. Follow testing checklist in template
  2. Verify all controls work
  3. Test negative cases (rejections)
  4. Update threat model after deployment

Document Deviations

If customizing beyond template:

// Custom rate limit for this endpoint
// Deviation from template: Increased to 10 requests per 15 min
// Reason: Users frequently update bio
withRateLimit(handler, { requests: 10, window: 15 * 60 })

Integration with Other Skills

Feature Implementation

1. Quick Implementation
   → security-prompts-controls (this skill)

2. Security Testing
   → prompt-engineering/08_security_testing.md

3. Update Threat Model
   → threat-modeling/08_update_model.md

Growing Complexity

1. Start Simple
   → security-prompts-controls (this skill)

2. Add Authorization
   → security-prompts-auth

3. Add Advanced Features
   → security-prompts-engineering

4. Review Security
   → security-prompts-threat-modeling

Related Skills

Parent Skill:

  • security-prompts - Main directory with all template categories (engineering, auth, threat-modeling, controls)

Implementation Skills (referenced by templates):

  • csrf-protection - CSRF implementation details
  • rate-limiting - Rate limiting patterns
  • input-validation - Validation strategies

Note: For other template categories (comprehensive implementation, authentication, threat modeling), see the parent security-prompts skill

Version History

v1.0 (2025-10-23): Initial skill creation

  • Converted 3 built-in-controls templates
  • Added comparison with prompt-engineering templates
  • Added upgrade paths
  • Integration with utility reference

Note: These templates are optimized for speed and simplicity. For comprehensive security guidance and complex features, see the parent security-prompts skill directory.

Quick Install

/plugin add https://github.com/harperaa/secure-claude-skills/tree/main/built-in-controls

Copy and paste this command in Claude Code to install this skill

GitHub 仓库

harperaa/secure-claude-skills
Path: security-prompts/built-in-controls

Related Skills

sglang

Meta

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

View skill

evaluating-llms-harness

Testing

This Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.

View skill

llamaguard

Other

LlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.

View skill

langchain

Meta

LangChain is a framework for building LLM applications using agents, chains, and RAG pipelines. It supports multiple LLM providers, offers 500+ integrations, and includes features like tool calling and memory management. Use it for rapid prototyping and deploying production systems like chatbots, autonomous agents, and question-answering services.

View skill