Skill
RetainCraft
Evidence-based AI-assisted learning protocol combining spaced repetition (SM-2), active recall, Feynman technique, interleaving, and deliberate practice. Fea...
When to use RetainCraft
Choose if
You're building an agent that coaches a human learner through a structured curriculum and want evidence-based methodology (SM-2 spaced repetition, active recall, Feynman technique, interleaving) plus burnout monitoring and L1-L5 mastery leveling. Best for exam prep or sustained technical-skill acquisition.
Avoid if
You need a one-shot quiz or flashcard tool — RetainCraft enforces ceremony (post-test recording, sequential leveling, two-consecutive-pass promotion). Also avoid when the learner won't reliably execute the Python `record-test` command after each module, since progress tracking degrades silently.
Risk Flags
- MEDIUM scope README states the system "relies on user compliance with Python command execution; skipped steps invalidate progress tracking". Without faithful `record-test` calls, level advancement is unreliable.
- LOW scope README caps simultaneous topics at three — exceeding "creates context fragmentation". Multi-domain learners must serialize.
- LOW runtime README requires Python 3.6+ and a bash shell; sandboxed JS-only runtimes won't be able to execute `scripts/srs.py`.
Cost
Type: Free
Dependencies
Minimum runtime: Python 3.6+
Distribution
- ClawHub
retaincraft- Repository
- https://github.com/kaixiad/RetainCraft
- License
- MIT