context-detection

The context-detection skill provides an intelligent, multi-layered mechanism for analyzing software repositories to identify the underlying technology stack, framework dependencies, and project structure. It is designed for developers, architects, and automated agents that need to establish a project's technical profile before executing tasks or loading environment-specific plugins. By analyzing configuration files, directory layouts, and current file contexts, this skill enables seamless integration of framework-specific capabilities, such as auto-loading testing helpers for Vitest, database patterns for Go, or state management tools for React.

• Multi-stack awareness: Expertly handles fullstack architectures where frontend (e.g., React, Vue) and backend (e.g., Go, Rust, Python) components coexist in the same repository.

• Hierarchical detection logic: Follows a strict priority order starting with explicit user settings in .claude/settings.json, followed by file extensions, configuration files (package.json, go.mod, Cargo.toml, pyproject.toml), and structural patterns.

• Skill discovery engine: Includes a robust discovery script to query official Claude Code skill locations, including project-specific, personal, marketplace, and enterprise directories.

• Adaptive framework support: Automatically maps detected technologies to appropriate development tools and skills, ensuring the AI agent is always equipped with the right specialized commands and context.

• Recursive structure analysis: Identifies directory patterns like src/routes or cmd/ to validate and confirm stack findings, providing higher confidence than configuration parsing alone.

• Detection reliability: When configuring projects, always ensure the priority hierarchy is respected; user overrides in .claude/settings.json will always supersede auto-detected findings.

• Performance and scope: The tool is optimized to ignore common dependency directories such as .git, node_modules, and vendor to maintain high-speed scanning.

• Multi-stack best practices: Use this skill specifically when working in monorepos or polyglot environments where different parts of the application require distinct development approaches.

• Input/Output: Receives the current working directory as an input; outputs a JSON payload containing identified stacks, associated framework skills, and summary metadata for the plugin orchestrator.