Science-backed plugins that make Claude Code work better.
Logos (λόγος) — reason, principle, rational order.
Every plugin in this marketplace is grounded in published research from cognitive science, behavioral economics, or software engineering. We don't guess what makes AI agents better — we apply what's already been proven in human performance research.
claude plugin marketplace add claude-code-logos- Evidence first — every mechanism cites published research
- Nudge, don't force — guide behavior through design, not restriction
- Measurable impact — each plugin should have a testable claim
Coming soon.
| Principle | Source | Application |
|---|---|---|
| Checklist effect | Gawande (2009); Hales & Pronovost (2006) | Task lists prevent step-skipping |
| Implementation intentions | Gollwitzer (1999) | Plan-then-execute improves completion rates |
| Cognitive load theory | Sweller (1988) | Break complex tasks into manageable chunks |
| Nudge theory | Thaler & Sunstein (2008) | Small design changes improve decisions |
| Deliberate practice | Ericsson et al. (1993) | Feedback loops improve performance |
| Error detection | Reason (1990) | Verification cycles catch mistakes |
MIT