Head-to-head pages
Sixteen side-by-side verdicts: what each one is good for, who should skip it, and one verifiable fact.
Anthropic Skills vs Addy Osmani Agent Skills
OpenDesign vs Diagram Design
fast-agent vs VexJoy Agent
MCPJam Inspector vs Guardana
Agent Skills (simota) vs Claude Skills (oaustegard)
Agent Skill Exchange vs EveryInfra SEO Skills
Awesome Agent Skills vs Agent Skill Exchange
Cairn Memory vs Hippo Memory
Melodic Plugins vs Lean Agents
OrchestralFlow vs chipping-orchestrator
Claude Skills by jsath vs oaustegard Claude Skills
Claude Skills by jsath vs Agent Skills (simota)
Melodic Plugins vs Universal Agent Plugins
motion-launch-videos vs Product Launch Motion
Universal Agent Plugins vs Awesome DSH Plugins- Melodic Plugins vs Awesome DSH Plugins
Side-by-side
Two or three entries, field by field. The list lives in the address bar, so a comparison is a link you can share — nothing is stored on this device.
Cells say not held when the catalogue has no value for that record — that is not a verdict on the project. A star count is shown only when GitHub reports one; we do not print a zero or a guess. The three usability rows come from the AI-tool record only (interface language, free tier, ICP filing); entries from other tables have no such columns and therefore say not held. Explore the catalogue: MCP servers · all entries.