How does it handle large codebases and context limits?

It automatically counts tokens for all files matched by your glob patterns, validates them against the model's limit, and reserves space for the generated response.

Is an API key required to use this skill?

Yes, you must provide an API key for your chosen provider (e.g., OPENAI_API_KEY) via environment variables or a command flag.

Which AI models can I use with the Consultant skill?

Consultant supports over 100 providers through LiteLLM, including OpenAI, Anthropic, Google Gemini, and local models via custom base URLs.

Can I run analyses in the background?

Yes, Consultant supports asynchronous execution with session management, allowing you to trigger a task and check its progress status later.

Do I need to install Python dependencies manually?

No, the skill uses uvx for automatic dependency management, installing required libraries like LiteLLM on the first run.

AI Architecture & Security Consultant

Name: AI Architecture & Security Consultant
Author: doodledood

bydoodledood

•

Security & Testing

Performs deep architectural reviews, security audits, and complex code analysis by leveraging 100+ AI models via LiteLLM.

About

The Consultant skill extends Claude's reasoning capabilities by connecting to specialized high-power AI models for demanding technical tasks. It streamlines complex workflows like full-codebase security audits, architectural debt analysis, and deep-dive logic reviews by handling asynchronous processing, token management across multiple files, and automatic model selection. Whether you need a second opinion from GPT-4o, Claude 3.5 Opus, or a local Llama instance, Consultant manages the session and context so you can focus on high-level decision-making.

Key Features

Automated token counting and context window overflow protection
Asynchronous session management for long-running, complex analyses
Cross-platform Python implementation with automatic dependency management via uvx
4 GitHub stars
Supports 100+ LLM providers including OpenAI, Anthropic, and local models via LiteLLM
Intelligent model discovery and selection based on task requirements

Use Cases

Onboarding to unfamiliar codebases by requesting high-level logic summaries
Generating comprehensive architectural reviews for complex system designs
Performing deep security audits to identify vulnerabilities across large repositories

About

Key Features

Automated token counting and context window overflow protection
Asynchronous session management for long-running, complex analyses
Cross-platform Python implementation with automatic dependency management via uvx
4 GitHub stars
Supports 100+ LLM providers including OpenAI, Anthropic, and local models via LiteLLM
Intelligent model discovery and selection based on task requirements

Use Cases

Onboarding to unfamiliar codebases by requesting high-level logic summaries
Generating comprehensive architectural reviews for complex system designs
Performing deep security audits to identify vulnerabilities across large repositories