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.
This skill delegates tech stack detection to a specialized lightweight agent, keeping your context lean.
Invoke this skill when ANY of these conditions are true:
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.
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
")
The agent will:
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
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.
This skill is particularly useful for:
init-explorer - Initial project understandingarchitecture-evaluator - Making architectural decisionstest-creator - Matching test framework conventionscoder - Following project patterns