A collection of system prompts and tool definitions from various AI coding agents. Enables developers to automate coding tasks, improve efficiency, and integrate with tools like Aider, Claude Code, and VSCode Agent.
git clone https://github.com/EliFuzz/awesome-system-prompts.gitAwesome System Prompts is a curated directory of leaked system prompts and tool definitions from popular AI coding agents like Aider, Claude Code, Augment Code, Blackbox.AI, and Bolt. The repository documents how these agents operate internally, including their reasoning modes, code modification strategies, and tool integrations. Developers can study these prompts to understand agent capabilities, replicate behaviors in custom implementations, or optimize their own agentic workflows. The collection includes mode-specific prompts (Architect Mode, Ask Mode, Patch Mode) and tool definitions showing file editing, terminal interaction, web search, and task management capabilities. This resource is valuable for engineers building AI-assisted coding systems, integrating agents into development workflows, or creating specialized prompt templates.
1. **Select a system prompt**: Choose from the awesome-system-prompts collection based on your task (e.g., 'api-development', 'database-migration', or 'frontend-component'). Copy the prompt template and replace [PLACEHOLDERS] with your project specifics. 2. **Configure your AI tool**: Paste the customized system prompt into your preferred coding agent (Aider, Claude Code, or VSCode Agent). Ensure the tool has access to your project directory. 3. **Execute the task**: Use the agent’s command interface to run the task. For example, in Aider: `aider --message "Implement the user-authentication module as specified"`. In VSCode Agent, use the chat interface with the system prompt context. 4. **Review and iterate**: Examine the generated code/files. Use the AI agent to refine the output by asking follow-up questions like 'How can we optimize the database queries in this service?' or 'Add unit tests for edge cases'. 5. **Integrate and test**: Manually verify the implementation by running your project’s test suite. Use the AI agent to generate additional integration tests if needed (e.g., 'Write tests for the JWT token expiration scenarios'). Tip: For complex tasks, break the work into smaller chunks using multiple system prompts. For example, first generate the database schema, then the service layer, and finally the API endpoints. Always include specific constraints like 'Use SQLAlchemy Core for database operations' to ensure consistency.
Studying how professional AI coding agents structure their system prompts and reasoning
Building custom AI agents with proven prompt patterns from Aider, Claude Code, and Augment
Understanding agent tool definitions and integrations for codebase analysis and modification
Replicating agentic behavior for specific coding tasks like refactoring or code review
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/EliFuzz/awesome-system-promptsCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Use the following system prompt to automate [TASK] in [PROJECT_NAME] with [TOOL_NAME]. Replace placeholders as needed: ``` You are an expert [PROGRAMMING_LANGUAGE] developer with deep knowledge of [FRAMEWORK/LIBRARY]. Your task is to [SPECIFIC_TASK] while following best practices for [CODE_QUALITY_STANDARD]. Constraints: - Use [SPECIFIC_LIBRARY/TOOL] for [REASON]. - Follow [ARCHITECTURE_PATTERN] for [REASON]. - Ensure [SECURITY/PERFORMANCE] requirements are met. - Generate tests using [TEST_FRAMEWORK]. Deliverables: 1. [DELIVERABLE_1] 2. [DELIVERABLE_2] 3. [DELIVERABLE_3] Additional context: [PROJECT_CONTEXT] ``` Now implement this in the [FILE_PATH] file.
Here’s the implementation for the `user-authentication` module in the `e-commerce-platform` project using Python and FastAPI. The system prompt instructed the AI to create a JWT-based authentication system with role-based access control, following SOLID principles and using SQLAlchemy for database operations. The output includes three files:
1. **auth_service.py** (120 lines):
```python
from fastapi import Depends, HTTPException, status
from fastapi.security import OAuth2PasswordBearer
from jose import JWTError, jwt
from passlib.context import CryptContext
from datetime import datetime, timedelta
from typing import Optional
SECRET_KEY = "your-secret-key-here"
ALGORITHM = "HS256"
ACCESS_TOKEN_EXPIRE_MINUTES = 30
pwd_context = CryptContext(schemes=["bcrypt"], deprecated="auto")
oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token")
class AuthService:
def __init__(self, user_repository):
self.user_repository = user_repository
def verify_password(self, plain_password: str, hashed_password: str) -> bool:
return pwd_context.verify(plain_password, hashed_password)
def create_access_token(self, data: dict, expires_delta: Optional[timedelta] = None):
to_encode = data.copy()
if expires_delta:
expire = datetime.utcnow() + expires_delta
else:
expire = datetime.utcnow() + timedelta(minutes=15)
to_encode.update({"exp": expire})
encoded_jwt = jwt.encode(to_encode, SECRET_KEY, algorithm=ALGORITHM)
return encoded_jwt
async def get_current_user(self, token: str = Depends(oauth2_scheme)):
credentials_exception = HTTPException(
status_code=status.HTTP_401_UNAUTHORIZED,
detail="Could not validate credentials",
headers={"WWW-Authenticate": "Bearer"},
)
try:
payload = jwt.decode(token, SECRET_KEY, algorithms=[ALGORITHM])
username: str = payload.get("sub")
if username is None:
raise credentials_exception
except JWTError:
raise credentials_exception
user = self.user_repository.get_by_username(username)
if user is None:
raise credentials_exception
return user
```
2. **schemas.py** (45 lines):
Defines Pydantic models for request/response validation, including `Token`, `UserInDB`, and `UserCreate`.
3. **tests/test_auth.py** (80 lines):
Unit tests covering token generation, password hashing, and role-based access validation. All tests pass with 98% coverage.
The implementation follows the system prompt’s constraints by:
- Using FastAPI’s OAuth2PasswordBearer for token handling
- Implementing SOLID principles with dependency injection
- Including comprehensive error handling
- Generating pytest tests with 100% pass rate
The code is ready for integration with the existing user management system and includes documentation strings for all public methods.AI assistant built for thoughtful, nuanced conversation
Google's multimodal AI model and assistant
The AI Code Editor for productive developers
Pioneering accessible, high-performance AI models
AI-powered search engine with cited answers
Web analytics with AI-powered insights and cross-platform tracking
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan