Research Synthesis

Research & Synthesis

Research Process

1. DEFINE    — What specific question am I trying to answer?
2. GATHER    — Collect sources (docs, benchmarks, community, real usage)
3. EVALUATE  — Assess source quality and recency
4. SYNTHESIZE — Extract patterns and insights across sources
5. DECIDE    — Apply findings to your specific context
6. DOCUMENT  — Record the decision and reasoning for future reference

---

Technology Evaluation Framework

For every technology comparison:

| Dimension | What to Evaluate | |---|---| | Fit | Does it solve your actual problem? | | Maturity | Production-proven or experimental? | | Community | Active maintenance, GitHub stars, Stack Overflow presence | | Performance | Benchmarks relevant to your use case | | Learning Curve | Team familiarity, documentation quality | | Ecosystem | Integrations, libraries, tooling | | Cost | License, infrastructure, ops overhead | | Exit Cost | How hard to migrate away if needed? |

---

Comparison Document Template

## Decision: [technology choice]

### Context
[What problem are we solving? What are our constraints?]

### Options Considered
| | Option A | Option B | Option C |
|--|---------|---------|---------|
| Fit | | | |
| Maturity | | | |
| Performance | | | |
| Team familiarity | | | |
| Long-term risk | | | |

### Decision
[What we chose and why]

### Trade-offs Accepted
[What we give up with this choice]

### Revisit If
[Conditions that would prompt reconsideration]

---

Source Quality Hierarchy

1. Official docs — authoritative, but may be biased toward positives
2. Benchmarks — verify the benchmark matches your use case
3. Production case studies — most reliable, but rare
4. Community discussions — useful for gotchas and real-world issues
5. Blog posts — variable quality, check date and author credentials

Always check the date. Technology moves fast — 2-year-old articles may be outdated.

---

Using AI for Research

"Research [topic] and summarize:
1. What problem it solves
2. When to use it vs alternatives [A, B, C]
3. Known limitations and gotchas
4. Current community sentiment (2025-2026)
5. One concrete example of production usage"

Cross-reference AI output with official docs — AI can hallucinate library APIs or outdated information.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: jamestorrevillas
• Source: jamestorrevillas/dev-skills
• License: MIT

Install and usage instructions live in the source repository linked above.