| name | research |
| description | Comprehensive technical research across GitHub, Google, and internal project knowledge. Use when tasked with finding libraries, researching implementation patterns, or looking up technical specifications. |
Research Skill
This skill provides a structured workflow for deep technical research. It leverages external search (Google, GitHub) and internal knowledge (codebase, documentation) to provide well-grounded answers and resources.
Workflow
When a research task is initiated, follow these steps:
- Define Search Goals: Identify the core technologies (e.g., Kotlin Multiplatform, Android, Compose) and the specific problem or feature.
- Internal Audit: Before going external, check if the codebase already has similar patterns or if
GEMINI.md / MEMORY.md contains relevant constraints or notes.
- External Search:
- Use
google_web_search for documentation, blogs, and broad queries.
- Use
web_fetch to analyze specific GitHub repositories, documentation pages, or issues.
- Refer to search_strategies.md for optimized query patterns.
- Synthesize and Validate: Combine findings into a coherent summary. If a library is found, verify its compatibility with the current project (check
libs.versions.toml, build.gradle.kts).
Research Tasks
1. Finding Libraries/Repositories
Use targeted GitHub searches via Google to find repositories. Look for stars, recent activity, and "Awesome" lists.
2. Implementation Patterns
Search for blog posts or official documentation from developer.android.com or kotlinlang.org. Focus on modern, idiomatic approaches (e.g., using Coroutines, StateFlow, Compose).
3. Debugging and Troubleshooting
Search GitHub issues and Stack Overflow for specific error messages or behavior descriptions.
Examples
- "Research how to implement biometric authentication in Compose Multiplatform."
- "Find a GitHub repo for a Kotlin Multiplatform QR code scanner."
- "Search for the latest best practices for Android security auditing."
Resource Navigation
