compare

name	compare
description	Side-by-side comparison of 2+ competing projects with stage-fair scoring. Supports mixing remote URLs and local paths. Produces a comparison matrix with ranked recommendation.

Research-Forge — Compare

Compare competing projects head-to-head across all analysis dimensions, judged stage-fairly.

Command: /research-forge:compare

Usage:

/research-forge:compare <target1> <target2> [target3...] [--lang zh|en]

Targets can be any mix of URLs, local directory paths, or local file paths.

Comparison Workflow:

1. Parallel Scan

• Run a scan (equivalent to /research-forge:scan) on each target in parallel using sub-agents.
• Collect metadata, metrics, stage classification, and first impressions for all projects.
• Adapt data collection strategy per target type (remote vs. local).

2. Comparison Matrix

Build a side-by-side table covering:

Dimension	Project A	Project B	Project C
Target Type	Remote/Local	Remote/Local	Remote/Local
Lifecycle Stage	S0-S3	S0-S3	S0-S3
Stars / Files
Contributors
Last Commit
License
Language
Business & Market
Problem Validity (1-5)
Why Now (1-5)
PMF Evidence (1-5)
Unit Economics (1-5)
Market Ceiling (1-5)
Growth Engine (1-5)
Technical & Architecture
Architecture Quality (1-5)
Engineering Culture (1-5)
Scalability (1-5)
Security Posture (1-5)
Documentation (1-5)
Maintainability & Tech Debt (1-5)
Moat & Strategy
Moat Strength (1-5)
Founder-Market Fit (1-5)
Org Capability (1-5)
Second Curve Readiness (1-5)
Overall Score (stage-weighted)

Mark any sub-dimension that is Context (×0) for a project's stage as ctx — it is shown for information but excluded from that project's Overall Score.

3. Stage-Fairness Rule

• Same stage → Overall Scores are directly comparable; rank by score.
• Different stages → raw Overall Scores are NOT directly comparable (each is weighted against different criteria). Compare each project against its own stage's One Question, then frame the recommendation as a maturity-vs-potential trade-off: "A is the safer, proven choice; B has the higher ceiling but is pre-PMF" — never as a naive score ranking.

4. Key Differentiators

For each project, identify:

• Unique Strength: What does this project do better than all others?
• Critical Weakness: What is the biggest gap vs. competitors?
• Best For: What use case or user profile is this project ideal for?

5. Verdict

• Rank projects from most to least recommended (respecting the stage-fairness rule).
• Provide context-dependent recommendation: "If you need X, choose A. If you need Y, choose B."
• Note each project's top Kill Criterion — the signal most likely to change its ranking.