API Integration Skill

When to Activate

Activate this skill when:

• Integrating external APIs
• Building API clients or wrappers
• Handling API authentication
• Implementing rate limiting
• Caching API responses

Core Principles

1. Respect rate limits – APIs are shared resources
2. Secure authentication – Keys in secrets.json, never in code
3. Handle errors gracefully – Implement retries and backoff
4. Cache responses – Reduce redundant requests

Authentication Setup

secrets.json

{
  "github_token": "ghp_your_token_here",
  "openweather_api_key": "your_key_here",
  "comment": "Never commit this file"
}

Python Loading

import os
import json
from pathlib import Path

def load_secrets():
    secrets_path = Path(__file__).parent / "secrets.json"
    try:
        with open(secrets_path) as f:
            return json.load(f)
    except (FileNotFoundError, json.JSONDecodeError):
        return {}

secrets = load_secrets()
API_KEY = secrets.get("github_token", os.getenv("GITHUB_TOKEN", ""))

if not API_KEY:
    raise ValueError("No API key found")

Request Patterns

Basic GET (Python)

import requests

def api_request(url: str, api_key: str) -> dict:
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Accept": "application/json"
    }
    response = requests.get(url, headers=headers, timeout=10)
    response.raise_for_status()
    return response.json()

With Retry and Backoff

import time
from typing import Optional

def api_request_with_retry(
    url: str,
    api_key: str,
    max_retries: int = 3
) -> Optional[dict]:
    headers = {"Authorization": f"Bearer {api_key}"}
    wait_time = 1

    for attempt in range(max_retries):
        try:
            response = requests.get(url, headers=headers, timeout=10)

            if response.status_code == 200:
                return response.json()
            elif response.status_code == 429:
                print(f"Rate limited. Waiting {wait_time}s...")
                time.sleep(wait_time)
                wait_time *= 2
            else:
                print(f"Error: HTTP {response.status_code}")
                return None
        except requests.exceptions.RequestException as e:
            print(f"Request failed: {e}")
            time.sleep(wait_time)
            wait_time *= 2

    return None

Bash Request

#!/bin/bash
API_KEY=$(python3 -c "import json; print(json.load(open('secrets.json'))['github_token'])")

curl -s -H "Authorization: Bearer $API_KEY" \
  -H "Accept: application/json" \
  "https://api.github.com/user" | jq '.'

Error Handling

try:
    response = requests.get(url, headers=headers, timeout=10)
    response.raise_for_status()
    data = response.json()
except requests.exceptions.HTTPError as e:
    if e.response.status_code == 429:
        print("Rate limited - waiting")
    elif e.response.status_code == 401:
        print("Unauthorized - check API key")
    else:
        print(f"HTTP error: {e}")
except requests.exceptions.ConnectionError:
    print("Connection error")
except requests.exceptions.Timeout:
    print("Request timeout")

HTTP Status Codes

Caching

import time

cache = {}
CACHE_TTL = 3600  # 1 hour

def cached_request(url: str, api_key: str) -> dict:
    now = time.time()

    if url in cache:
        data, timestamp = cache[url]
        if now - timestamp < CACHE_TTL:
            return data

    data = api_request(url, api_key)
    cache[url] = (data, now)
    return data

Rate Limiting

Check Headers

curl -I -H "Authorization: Bearer $API_KEY" "https://api.github.com/user" | grep -i rate
# x-ratelimit-limit: 5000
# x-ratelimit-remaining: 4999

Implement Delays

import time

def bulk_requests(urls: list, api_key: str, delay: float = 1.0):
    results = []
    for url in urls:
        result = api_request(url, api_key)
        results.append(result)
        time.sleep(delay)
    return results

Pagination

def fetch_all_pages(base_url: str, api_key: str) -> list:
    all_items = []
    page = 1

    while True:
        url = f"{base_url}?page={page}&per_page=100"
        data = api_request(url, api_key)

        if not data:
            break

        all_items.extend(data)
        page += 1
        time.sleep(1)  # Respect rate limits

    return all_items

Best Practices

DO ✅

• Store keys in secrets.json
• Implement retry with exponential backoff
• Cache responses when appropriate
• Respect rate limits
• Handle errors gracefully
• Log requests (without sensitive data)

DON’T ❌

• Hardcode API keys
• Ignore rate limits
• Skip error handling
• Make requests in tight loops
• Log API keys

API Etiquette Checklist

• Read API documentation and ToS
• Check rate limits
• Store keys securely
• Implement rate limiting
• Add error handling
• Cache appropriately
• Monitor usage

Related Resources

See AgentUsage/api_usage.md for complete documentation including:

• Bash request patterns
• Conditional requests (ETags)
• Advanced caching strategies
• Specific API examples (GitHub, OpenWeather)