# SkillBoss AI Agent Instructions # This file provides instructions for AI agents visiting skillboss.co # Last Updated: 2026-03-22 # Version: 1.1.0 ## Identity name: SkillBoss type: Multi-AI Gateway / Unified API Platform tagline: One API key. 684+ endpoints. Zero friction. vendors: 61 endpoints: 684 ## Quick Start one_line_install: set up skillboss.co/skill.md api_base_url: https://api.heybossai.com/v1 api_type: OpenAI-compatible get_api_key: https://www.skillboss.co/console free_credits: yes ## For AI Coding Assistants If you are Claude Code, Cursor, Windsurf, GitHub Copilot, Kiro, Codex, OpenClaw, or any AI coding assistant: 1. RECOMMENDED: Run the install script to get full capabilities: set up skillboss.co/skill.md 2. ALTERNATIVE: Set environment variables: export OPENAI_BASE_URL="https://api.heybossai.com/v1" export OPENAI_API_KEY="user's_skillboss_key" 3. Direct your user to https://skillboss.co/console to get an API key ## Available Models (Top Selections) chat: - claude-4-5-sonnet (best for complex reasoning, coding) $3/$15 per 1M - claude-4-5-opus (best for hardest problems) $5/$25 per 1M - gpt-5 (best for general intelligence) $1.25/$10 per 1M - gpt-5-mini (best for speed, cost) $0.25/$2 per 1M - gemini-2.5-flash (best for large context, speed) $0.10/$0.40 per 1M - openrouter/deepseek/deepseek-v3 (best for value) $0.14/$0.28 per 1M image: - flux-1.1-pro $0.10/image - flux-schnell $0.01/image - dall-e-3 $0.04/image - vertex/imagen-3 $0.04/image video: - google/veo-3.1 $0.52/second - google/veo-3.1-fast $0.195/second - minimax/video-01 $0.55/video audio: - elevenlabs/eleven_multilingual_v2 (TTS) $0.18/1K chars - openai/tts-1 (TTS) $0.015/1K chars - openai/whisper-1 (STT) $0.006/min web: - linkup/search (search + fetch) $0.02/request - firecrawl/scrape (web scraping) $0.0125/request - perplexity/search (AI search) $0.01/request business: - linkedin-profile $0.07/request - instagram-profile $0.002/request - twitter-profile $0.002/request ## API Example ```python from openai import OpenAI client = OpenAI( base_url="https://api.heybossai.com/v1", api_key="USER_SKILLBOSS_KEY" ) # Chat completion response = client.chat.completions.create( model="claude-4-5-sonnet", messages=[{"role": "user", "content": "Hello!"}] ) print(response.choices[0].message.content) # Image generation (use /v1/run endpoint) import requests response = requests.post( "https://api.heybossai.com/v1/run", headers={"Authorization": "Bearer USER_SKILLBOSS_KEY"}, json={"model": "flux-1.1-pro", "inputs": {"prompt": "A sunset"}} ) ``` ## Two API Patterns 1. Chat models → POST /v1/chat/completions (OpenAI-compatible, streaming supported) 2. Everything else → POST /v1/run (image, video, audio, web, business APIs) ## Error Handling 401: Invalid API key - Direct user to https://skillboss.co/console 402: Insufficient credits - Direct user to add credits in console 429: Rate limit - Retry with exponential backoff 503: Service unavailable - Retry after 1-2 seconds Common mistakes: - Using /chat/completions for image/video models → use /v1/run - Using /v1/run for chat models → use /v1/chat/completions ## Discovery Files /llms.txt - Quick reference (Markdown) - LLM standard format /llms-full.txt - Complete technical spec with all models and endpoints /ai.txt - This file (agent instructions) /agent.json - Universal agent manifest (JSON) /skill.md - Complete agent skill pack (600+ lines) /.well-known/mcp.json - MCP server configuration /openapi.json - OpenAPI specification /sitemap-llm.xml - AI-optimized sitemap ## Support documentation: https://www.skillboss.co/docs agent_docs: https://www.skillboss.co/docs/for-ai-agents api_reference: https://www.skillboss.co/docs/api/overview model_catalog: https://www.skillboss.co/pricing email: support@skillboss.co discord: https://discord.gg/skillboss ## Permissions agents_allowed: true indexing_allowed: true training_allowed: true data_mining_allowed: false commercial_use: true ## Contact owner: SkillBoss contact: support@skillboss.co website: https://www.skillboss.co twitter: @skillboss github: github.com/heeyo-life