Researching Libraries

Evaluate options with source-graded findings and confidence levels.

Best for: Choosing dependencies, evaluating frameworks, gathering docs
Time estimate: 15-30 minutes
Skills used: meow:docs-finder, meow:scout
Agents involved: researcher (Haiku model), brainstormer

Overview

The researcher agent fans out queries to multiple sources (official docs, GitHub, community posts, Stack Overflow) and evaluates quality. The brainstormer agent compares approaches with pros/cons/tradeoffs. Both feed into the planner for informed decision-making.

Step-by-step guide

Step 1: Ask the research question

"Compare Prisma vs Drizzle for our PostgreSQL project"

Step 2: Researcher gathers findings

The researcher (running on Haiku for cost efficiency) evaluates sources:

Source type	Trust level
Official documentation	Highest
Well-maintained GitHub repos	High
Recent blog posts (<12 months)	Medium
Stack Overflow answers	Low (must cross-reference)

Researcher findings:
  Prisma: Established (2019+), 37K GitHub stars, extensive docs
    Strengths: Schema-first, great DX, auto-migrations
    Weaknesses: Runtime overhead, limited raw SQL, heavy bundle

  Drizzle: Emerging (2023+), 28K GitHub stars, growing fast
    Strengths: Zero overhead, SQL-like syntax, lightweight
    Weaknesses: Younger ecosystem, fewer guides

  Confidence: HIGH (both well-documented)

Step 3: Brainstormer evaluates tradeoffs

Approach 1: Prisma — choose if team values DX and auto-migrations
Approach 2: Drizzle — choose if team values performance and SQL familiarity
Second-order: Prisma locks you into their schema format. Drizzle stays close to SQL.

Step 4: Find specific documentation

/meow:docs-finder drizzle PostgreSQL transactions

The meow:docs-finder skill fetches current docs via Context7 or Context Hub, avoiding stale training data.

Next workflow

→ Maintaining Old Projects — work in unfamiliar codebases