Skip to content
Back to skills

Reinforcement

ASecurity

Turn every gap (wrong or poorly reasoned answer) into a targeted reinforcement note and a retest until the gap is closed. Use after a quiz/test/review opens gaps, or when open gaps exist on the active topic.

  • 3 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 18, 2026
ai-agentsgo

Security analysis

A100/100

Scanned November 18, 2026

npx -y skills add Lelpoco2/learning-curve --skill reinforcement --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Reinforcement?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Reinforcement
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/lelpoco2-reinforcement/badge)](https://www.skillsdirectory.com/skills/lelpoco2-reinforcement)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: reinforcement
description: Turn every gap (wrong or poorly reasoned answer) into a targeted reinforcement note and a retest until the gap is closed. Use after a quiz/test/review opens gaps, or when open gaps exist on the active topic.
---

# Reinforcement

A wrong answer means a piece of reasoning is not consolidated. The goal is to find that
piece and rebuild it — not to repeat the lesson.

## For each open gap

1. Diagnose: re-read the learner's answer and reasoning. Name the exact misconception
   (e.g. "treats the dot product as a vector", "adds velocities in the rotating frame").
   If unclear, ask one short question.
2. Write the reinforcement note at the path given by `quiz_record` (or
   `learning_state` paths with `lesson` and `concept`), from `reinforcement.md`:
   - What happened: quote the answer, point to the exact step.
   - Why that reasoning is tempting: what is right about it, where it stops working.
   - Rebuild the idea from a DIFFERENT angle than the lesson (Learning Style summary).
   - One memory aid aimed at this exact mistake (skill `memory-aids`): a sign error → a
     special-value check; two swapped formulas → a contrast pair; a forgotten hypothesis →
     the counterexample that shows why it is needed.
   - Side by side: wrong vs right reasoning.
   - 2 practice questions with folded answers.
   Short: 1 screen. Verify math with `math_check` and facts with `source_lookup`
   (if the learner's answer follows their textbook's convention, it may not be wrong). Link it from the lesson note
   (a `> [!warning]` callout near the relevant section: "Common gap: … → [[link]]").
3. Walk through it in chat briefly (4–8 lines), then ask if it is clear.
4. Retest: 1–3 NEW questions on the concept (`quiz_ask`, independent, askWhy).
   Grade, then `quiz_record` with kind `retest` and the same concept names.
   - Correct with sound reasoning → the gap closes (tool reports it). Set the
     reinforcement note `status: resolved`.
   - Still wrong → try another angle (consider a missing prerequisite), update the note,
     retest again later in the session or next session.
5. Record which angle worked with `learner_profile` observe.

## Order

Close gaps of the current block before moving to the next block when possible; never
let more than ~5 gaps accumulate. Gaps from a subject final exam belong to the topic of
the question: write the note in that topic's lesson folder and retest with
`quiz_record` kind `retest`, `subject` and `topic` set (closing the last one completes
the exam). Gaps found in reviews follow the same procedure but
the note can be shorter.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.