tech-stack-analyzer

leonj1's avatarfrom leonj1

Delegates tech stack detection to a lightweight agent. Use when you need to know the project's language, frameworks, or dependencies without loading manifest files into context.

⭐0stars🔀0forks📁View on GitHub🕐Updated Dec 21, 2025
[Developer Tools]

When & Why to Use This Skill

The Tech Stack Analyzer is a specialized Claude skill designed to identify a project's programming languages, frameworks, and dependencies efficiently. By delegating analysis to a lightweight sub-agent, it provides a concise summary of the development environment while reducing context window usage by up to 90%, making it an essential tool for large-scale codebase exploration and architectural assessment.

Use Cases

  • Initial Project Onboarding: Quickly understand the core technologies and structure of a new repository without manual file inspection.
  • Context Window Optimization: Analyze large manifest files like package.json or requirements.txt without exhausting the AI's token limit.
  • Architectural Consistency: Identify existing patterns and libraries to ensure new code or services align with the project's established tech stack.
  • Automated Test & Build Setup: Detect configured testing frameworks and build tools to generate compatible scripts and configurations.
nametech-stack-analyzer
descriptionDelegates tech stack detection to a lightweight agent. Use when you need to know the project's language, frameworks, or dependencies without loading manifest files into context.

Tech Stack Analyzer Skill

This skill delegates tech stack detection to a specialized lightweight agent, keeping your context lean.

When to Invoke This Skill

Invoke this skill when ANY of these conditions are true:

  1. Need to know the tech stack: You need to identify the project's language, framework, or dependencies
  2. Framework-specific decisions: You need to choose patterns or libraries based on what the project uses
  3. Test framework identification: You need to know what testing framework is configured
  4. Build tool detection: You need to know how to build, run, or test the project
  5. Dependency analysis: You need to check what packages/libraries are already installed

Why Use This Skill?

Without this skill: You would read package.json, go.mod, requirements.txt, etc. directly, loading hundreds of lines of dependencies into your context.

With this skill: The tech-stack-analyzer agent (haiku model) reads manifests and returns a concise 30-50 line summary.

Context savings: 70-90% reduction in manifest-related context usage.

Invocation

When you need tech stack information, invoke the agent:

Task(subagent_type="tech-stack-analyzer", prompt="
Analyze this project's tech stack.
")

For specific focus areas:

Task(subagent_type="tech-stack-analyzer", prompt="
Analyze this project's tech stack.
Focus on: testing framework and test utilities
")

What tech-stack-analyzer Will Do

The agent will:

  1. Find manifests: Locate package.json, go.mod, requirements.txt, Cargo.toml, etc.
  2. Read selectively: Read only first 50-100 lines of each manifest
  3. Extract key info: Language, runtime, frameworks, test tools, build tools
  4. Detect patterns: Monorepo, microservices, serverless structures
  5. Return summary: Concise structured output with versions and key dependencies

Expected Output

You will receive a structured summary like:

## Tech Stack Summary

**Language**: TypeScript 5.3
**Runtime**: Node.js 20.x

**Frameworks**:
- Web: Express 4.18
- UI: React 18.2

**Testing**:
- Framework: Jest 29
- Utilities: React Testing Library

**Build**:
- Bundler: Vite 5.0
- Task Runner: npm scripts

**Key Dependencies**:
- Database: PostgreSQL (pg driver)
- ORM: Prisma 5.7

Example Usage

Scenario: You need to write a new service and want to match existing patterns.

Without skill: Read package.json (200+ lines), tsconfig.json, check for frameworks manually.

With skill:

Task(subagent_type="tech-stack-analyzer", prompt="Analyze this project's tech stack.")

Result: You know it's Express + TypeScript + Prisma in 30 lines of context.

Do NOT Invoke When

  • You already know the tech stack from earlier in the conversation
  • The user has explicitly told you the tech stack
  • You only need to check a single specific file (use Read directly)
  • The project has no manifest files (empty/new project)

Consumers

This skill is particularly useful for:

  • init-explorer - Initial project understanding
  • architecture-evaluator - Making architectural decisions
  • test-creator - Matching test framework conventions
  • coder - Following project patterns