跳转至

course-delivery-review

所属包: course

Close the loop after course delivery. Use when reviewing teaching feedback,

兼容性: Designed for OpenCode. Assumes repo-local .opencode/skills discovery

Purpose

Turn post-delivery feedback into diagnosed issues and a concrete course iteration plan, closing the loop from teaching back into content authoring.

Inputs

Collect whatever is available; do not block on missing channels:

  • learner feedback (surveys, interviews, informal comments)
  • quiz/assessment score distributions per lesson or module
  • completion rate (per module, per lab, overall)
  • frequent Q&A questions (答疑高频问题) from delivery sessions

Diagnostic categories

Classify each finding into one primary category:

  • outdated content (内容过时): facts, tools, or versions no longer current
  • difficulty mismatch (难度错配): content too hard or too easy for the actual audience baseline
  • ineffective exercise (练习失效): labs/quizzes that learners fail for reasons unrelated to the learning objective, or complete without learning
  • duration drift (时长偏差): lessons or labs that consistently overrun or underrun their allocated time

Mapping to the QC status model

Reuse the pack QA/QC conventions; do not invent a parallel model:

  • severity: blocker / major / minor (e.g. wrong technical fact = blocker; systematic difficulty mismatch in a core module = major; single-lesson duration drift = minor)
  • status: open / in-progress / fixed / verified / deferred (see the QC status model used by course-quality-control-reporting)

Each diagnosed issue gets: category, severity, status, evidence (which input channel surfaced it), and scope (module/lesson).

Default output structure

Write the iteration report to docs/07-qc/ (delivery review is a QC-loop artifact):

# Delivery Review <course> <date/round>

## Input summary
## Diagnosed issues (category / severity / status / evidence / scope)
## Iteration plan (which module, which lesson, what to change)
## Deferred items and rationale
## Next review checkpoint

The iteration plan must name concrete targets: module, lesson, and the specific change (rewrite section, replace lab, adjust duration budget, update outdated references).

Feeding the revision loop

  • Route content changes to course-content-authoring; route outline-level changes (module structure, hour reallocation) to course-outline-design first.
  • Route lab/exercise fixes to course-lab-design.
  • After revision, record status transitions and run QA/QC before any re-release; the delivery review report feeds the next QC round as findings input.

Gotchas

  • Do not treat one vocal complaint as a pattern; require evidence from at least one quantitative channel (scores, completion) or repeated qualitative signals.
  • Do not diagnose from completion rate alone; a low rate can mean difficulty mismatch, duration drift, or logistics — separate the hypotheses.
  • Do not bypass QA/QC: iteration suggestions are inputs to authoring, not direct edits to release material.
  • Do not lose severity/status when converting feedback into issues.