agentsclimarketplace

Course retro

Skill DojoCodingLabs/instructional-design-toolkit/skills/course-retro

Ingiere xAPI + feedback CSV de una cohorte y produce retro report Kirkpatrick L1-L4 con iteration candidates priorizados (impact × ease). Use when user asks to "course retrospective", "retro", "post-mortem cohort", "Kirkpatrick analysis", "analizar cohort", "retro del curso", "/course-retro".From its SKILL.md

Install
npx -y skills add DojoCodingLabs/instructional-design-toolkit --skill course-retro

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.6 KB, as published. Nobody here has run it

Retro de Cohort

Análisis post-cohort: ingesta xAPI statements + feedback forms, mapea a Kirkpatrick L1-L4, prioriza qué iterar.

Inputs

  • course-slug (required).

Flujo

Paso 1 — Load course + cohort number

  1. Cargar course.json.
  2. Detectar siguiente N para retros/retro-cohort-{N}.md (incrementar desde el último).

Paso 2 — Ingest xAPI

"¿Tenés un CSV con xAPI statements de la cohort? (ver formato esperado en ${CLAUDE_PLUGIN_ROOT}/references/adapters/xapi/csv.md).

Path al archivo (o 'skip' si no tenés):"

Si provisto: leer y parsear según adapters/xapi/csv.md. Si skip: WARN "L2 Learning no podrá ser medido sin xAPI data" y continuar.

Paso 3 — Ingest feedback (Tally/Typeform)

"¿Tenés CSVs de Tally o Typeform por módulo? (uno por módulo, formato documentado en ${CLAUDE_PLUGIN_ROOT}/references/adapters/feedback/).

Lista de paths (uno por línea, o 'skip'):"

Si provistos: parsear según adapters/feedback/{tally,typeform}-csv.md. Si skip: WARN "L1 Reaction no podrá ser medido sin feedback data" y continuar.

Paso 4 — L3 + L4 manual prompts (opcional)

"Kirkpatrick L3 (Behavior 30 días después) y L4 (Results / capstone tracking) requieren input manual.

L3: ¿hiciste follow-up survey 30 días post-curso? Si sí, dame los datos clave (ej. '5/8 siguen usando Riverpod en proyectos personales').

L4: ¿cuántos alumnos completaron el capstone? ¿Con qué calidad?"

Paso 5 — Mapear a Kirkpatrick L1-L4

Por cada level, agregar la data:

  • L1 Reaction: por módulo — avg rating, top 3 feedback recurrentes (free text cluster), confusion points.
  • L2 Learning: por AU — pass rate (passed / launched), score distribution, time-on-task vs estimated.
  • L3 Behavior (si data manual provista): summary qualitativo.
  • L4 Results (si data manual provista): capstone completion rate + quality highlights/lowlights.

Paso 6 — Priorize iteration candidates

Para cada signal anómalo (rating < 3.5, pass rate < 60%, time 2x estimated, etc.), generar un iteration candidate con:

  • Descripción del problema
  • Impact estimate (cuántos alumnos afectados)
  • Ease estimate (1-5: trivial → architectural change)
  • Score = impact × ease

Ordenar por score descendente. Top 5 al usuario.

Paso 7 — Generate report

docs/instructional-design/courses/{slug}/retros/retro-cohort-{N}.md:

# Retro: {course title}, Cohort {N}
**Date**: {today}
**Cohort size**: {N learners}
**Course version at cohort start**: {version}

## L1 Reaction
| Módulo | Rating avg | N | Top feedback |
|---|---|---|---|

## L2 Learning
| AU | Pass rate | Score distribution | Time vs estimate |
|---|---|---|---|

## L3 Behavior (manual)
[summary]

## L4 Results (manual)
[summary]

## Top 5 Iteration Candidates (impact × ease)
1. ...

PUERTA (sola): confirmar antes de escribir el report.

Recursos

  • ${CLAUDE_PLUGIN_ROOT}/references/adapters/xapi/csv.md
  • ${CLAUDE_PLUGIN_ROOT}/references/adapters/feedback/tally-csv.md
  • ${CLAUDE_PLUGIN_ROOT}/references/adapters/feedback/typeform-csv.md
  • ${CLAUDE_PLUGIN_ROOT}/assets/skill-references/kirkpatrick-evaluation.md
  • ${CLAUDE_PLUGIN_ROOT}/references/kirkpatrick-feedback-tools.md

Keep looking

Skills are one crate of 609,689. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.