What are the different thinking modes in Spec Tech Research?

The skill offers four modes: 'think' (10k tokens) for overviews, 'think-hard' (20k) for comparisons, 'think-harder' (30k) for complex decisions, and 'ultrathink' (50k) for maximum depth and academic analysis.

Can I export the research reports to specific file formats?

Yes, you can use the --output (-o) flag to specify a filename and the --format (-f) flag to choose between Markdown or JSON output structures.

What kind of report templates are available?

Standard templates include 'overview' for summaries, 'comprehensive' for full analysis, 'academic' for formal research with citations, and 'technical' for implementation-focused guides with code.

How does this skill use my local codebase context?

It leverages the Serena MCP to search for specific implementation patterns and symbols in your repository, allowing the research results to be architecture-aware and relevant to your existing code.

Spec Tech Research

Name: Spec Tech Research
Author: sc30gsw

bysc30gsw

•

학습 및 문서화

Conducts deep technical analysis and generates structured research reports using advanced thinking modes and codebase-aware MCP integration.

The Spec Tech Research skill transforms Claude into a high-powered technical analyst capable of performing exhaustive domain research. By utilizing graduated 'Thinking Modes' with token budgets up to 50,000, it enables Claude to tackle complex architectural decisions and academic-level technical investigations. The skill prioritizes Model Context Protocol (MCP) integrations like Serena for real-time codebase intelligence, ensuring that research findings are tailored to your specific implementation patterns. Whether you need a quick framework comparison or a 10-page technical specification, this skill provides the structure and depth required for professional software engineering.

주요 기능

01Automated diagram generation and citation support for verified research findings

028 GitHub stars

03Deep integration with Serena MCP for codebase-aware pattern matching and symbol analysis

04Four graduated thinking modes (up to 50,000 tokens) for multi-layered analysis

05Customizable reporting templates including Academic, Technical, and Comprehensive formats

06Configurable confidence thresholds and multi-language output capabilities

사용 사례

01Generating comprehensive technical documentation and implementation guides for new library integrations

02Comparing architectural patterns like Microservices vs. Monoliths within the context of an existing codebase

03Performing deep-dive security research and implementation analysis for complex protocols like OAuth 2.0

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sc30gsw/claude-code-customes spec-tech-research

For use in Claude.ai and ChatGPT

Download Skill