Technology Integration Research
A focused online research methodology for answering specific technical questions about how tools, frameworks, and services work together.
When to Use
Use this skill when you need to:
- •Understand how Tool A integrates with Tool B
- •Find migration patterns from one technology to another
- •Discover configuration requirements and setup procedures
- •Learn about API usage and authentication flows
- •Identify common pitfalls and recommended approaches
- •Research current best practices (2024-2026)
Key indicator: Questions containing "How does X work with Y?" or "What do I need to set up Z?"
Research Methodology
1. Context Gathering
- •Check project documentation for current technology stack
- •Note versions of tools already in use
- •Understand what's changing (migration, new integration, upgrade)
- •Identify specific question to answer
2. Search Strategy
Official Documentation First:
- •Tool A official docs
- •Tool B official docs
- •Official integration guides (if they exist)
Current Practices:
- •Recent blog posts (2024-2026)
- •GitHub issues and discussions
- •Migration guides and changelogs
- •Example repositories or starter templates
Search Query Patterns:
code
"[Tool A] [Tool B] integration 2026" "[Tool A] with [Tool B] setup guide" "migrate from [Old Tool] to [New Tool] 2026" "[Tool A] authentication [specific auth system]" "[Tool A] Docker Compose example" "[Tool A] version [X.x] [Tool B]"
3. Source Validation (Ongoing)
Prioritization:
- •Official documentation (highest authority)
- •Official blog posts from project maintainers
- •Well-maintained community guides with recent updates
- •GitHub discussions with maintainer responses
- •Individual blog posts (verify date and version)
Confidence Levels: Assign confidence to each finding based on source quality and convergence:
- •High: Official docs confirm, or 3+ independent sources agree
- •Medium: 1-2 reliable sources, or official docs are ambiguous
- •Low: Single community source, or sources conflict
Red Flags:
- •Information older than 2 years (unless researching legacy)
- •Blog posts without version numbers
- •Contradictory information without explanation
- •Examples that don't show imports/dependencies
- •Tutorials that skip authentication/security
4. Information Synthesis (5 minutes)
Extract and organize:
- •Key integration patterns discovered
- •Required configuration (env vars, config files)
- •Dependencies to install
- •Step-by-step setup if available
- •Common pitfalls mentioned across sources
- •Gaps in documentation or open questions
Output Format
Structure findings for easy scanning and comparison:
markdown
## Question Researched [Restate the specific question clearly] ## Summary [2-3 sentence overview of key findings] ## Key Findings - **[High/Medium/Low]** Finding 1: [Concise statement with source] - **[High/Medium/Low]** Finding 2: [Concise statement with source] - **[High/Medium/Low]** Finding 3: [Concise statement with source] ## Official Documentation - [Tool Name Docs](URL) - What it covers - [Integration Guide](URL) - What it covers ## Integration Pattern [Code snippet or config example showing the integration] ```language // Example code from official docs ```
Setup Requirements
Dependencies:
- •package-name@version - What it's for
Environment Variables:
- •
VAR_NAME- Purpose and example value
Services Needed:
- •Service name - How it's used
Configuration Example
[Actual config file example if found]
Common Pitfalls
- •Issue 1: What goes wrong and how to avoid it
- •Issue 2: What goes wrong and how to avoid it
Version Compatibility
- •Tool A: version range tested
- •Tool B: version range tested
- •Known issues: [list any]
Open Questions / Gaps
- •What couldn't be found or remains unclear
- •What needs further investigation
Recommended Next Steps
- •[Concrete action to take based on findings]
- •[Next action]
- •[etc.]
Sources Consulted
- •Source Title - Published/Updated: [date], Accessed: [date]
- •Source Title - Published/Updated: [date], Accessed: [date]
- •[etc.]
code
## Quality Checklist Before returning findings: - [ ] Checked 3+ sources minimum - [ ] Included at least one official source - [ ] Assigned confidence levels (High/Medium/Low) to each finding - [ ] Noted any conflicts between sources - [ ] Included code/config examples (not just prose) - [ ] Verified information is current (2024-2026) - [ ] Listed all sources with URLs and publication dates - [ ] Identified gaps or unclear areas - [ ] Included concrete recommended next steps - [ ] Structured output in standard format ## Example Questions **Good (specific and actionable):** - "How does Tool A authenticate with Tool B's JWKS endpoint?" - "What's the recommended Docker Compose setup for Service X with Database Y?" - "How do I migrate from Library A to Library B?" **Too Broad (needs scoping):** - "How does Tool X work?" → Scope to specific aspect - "Tell me about Technology Y" → What specifically about it? - "Best practices for category Z" → Too vague, narrow the question ## Scaling Effort to Question Complexity Match your research effort to the question's complexity: **Simple question** (e.g., "What package integrates X with Y?"): - Check official docs for both tools - Find 2-3 community examples or guides - Stop when sources converge on the same answer **Moderate question** (e.g., "How do I set up X with Y and Z?"): - Check official docs for all tools involved - Find 5-8 sources across different angles (setup guides, examples, discussions) - Look for complete working examples - Stop when you have a clear integration pattern and setup steps **Complex question** (e.g., "How do I migrate from A to B while maintaining C?"): - Check official migration guides and docs - Find 10+ sources including GitHub discussions, migration stories, and examples - Look for edge cases and common pitfalls - Compare multiple approaches if they exist - Stop when you have clear recommendations with trade-offs explained **Too complex?** If you've checked 15+ sources and still need more depth, or the question involves architectural decisions beyond simple integration, consider using the investigator agent instead.