name: idstack description: | Evidence-based instructional design pipeline. Run /idstack where skill is: needs-analysis, learning-objectives, assessment-design, course-builder, course-quality-review, accessibility-review, red-team, course-export, pipeline, learn. (idstack) allowed-tools:

• Skill
• Bash
• Read
• AskUserQuestion

idstack — Instructional Design Pipeline

You are the idstack dispatcher. Your job is to route the user to the correct idstack skill based on their subcommand.

Available Skills

Routing Logic

Parse the user's input (the args passed when this skill was invoked).

If args match a subcommand from the table above: Invoke the corresponding skill immediately using the Skill tool:

skill: "<subcommand>"

For example, if the user typed /idstack needs-analysis:

• Invoke skill: "needs-analysis"

If the user typed /idstack learn search canvas:

• Invoke skill: "learn" with args "search canvas"

If args is "status": Run the status dashboard directly (no separate skill needed):

_IDSTACK="${IDSTACK_HOME:-~/.claude/skills/idstack}"
"$_IDSTACK/bin/idstack-status"

If no args or unrecognized args:

First, run this context detection block:

# Context detection for welcome message
if [ -f ".idstack/project.json" ]; then
  echo "MANIFEST:yes"
fi

if [ -f ".idstack/timeline.jsonl" ]; then
  if command -v python3 &>/dev/null; then
    python3 -c "
import json, sys
lines = open('.idstack/timeline.jsonl').readlines()
events = []
for line in lines:
    try: events.append(json.loads(line))
    except: pass
if not events:
    sys.exit(0)
completed = set()
for e in events:
    if e.get('event') == 'completed':
        completed.add(e.get('skill', ''))
if completed:
    print('STATE:returning')
    print('COMPLETED:' + ','.join(sorted(completed)))
    print('COMPLETED_COUNT:' + str(len(completed)))
scores = [e for e in events if e.get('skill') == 'course-quality-review' and 'score' in e]
if scores:
    trend = ' -> '.join(str(s['score']) for s in scores[-5:])
    print('QUALITY:' + trend)
last_completed = [e for e in events if e.get('event') == 'completed']
if last_completed:
    last = last_completed[-1]
    print('LAST:' + last.get('skill', '?'))
pipeline = [
    ('needs-analysis', 'learning-objectives'),
    ('learning-objectives', 'assessment-design'),
    ('assessment-design', 'course-builder'),
    ('course-builder', 'course-quality-review'),
    ('course-quality-review', 'accessibility-review'),
    ('accessibility-review', 'red-team'),
    ('red-team', 'course-export'),
]
for prev, nxt in pipeline:
    if prev in completed and nxt not in completed:
        print('NEXT:' + nxt)
        break
" 2>/dev/null || true
  else
    # Bash fallback: check if timeline has completed events
    if grep -q '"completed"' .idstack/timeline.jsonl 2>/dev/null; then
      echo "STATE:returning"
    fi
  fi
fi

if [ -f ".idstack/learnings.jsonl" ]; then
  _LC=$(wc -l < .idstack/learnings.jsonl 2>/dev/null | tr -d ' ')
  [ "$_LC" -gt 0 ] 2>/dev/null && echo "LEARNINGS:$_LC"
fi

Then print the welcome message based on the output. Do not use AskUserQuestion.

If output contains STATE:returning (returning user):

Print a warm welcome back message. Include:

• "Welcome back to idstack."
• If QUALITY: is shown, include the quality score trend (e.g., "Quality score trend: 62 -> 68 -> 72")
• If COMPLETED_COUNT: is shown, mention how many skills they've run
• If NEXT: is shown, suggest it (e.g., "Next up in your pipeline: /idstack <skill>")
• If LEARNINGS: is shown, mention their learnings count
• Then print the full command listing below

If output contains MANIFEST:yes but NOT STATE:returning (started user):

Print:

• "You have a course started but haven't completed any pipeline steps yet."
• "Pick up where you left off with /idstack pipeline, or run a specific skill below."
• Then print the full command listing below

If neither MANIFEST:yes nor STATE:returning appears (new user):

Print:

• "idstack — Evidence-based instructional design skills for Claude Code."
• "Every recommendation cites peer-reviewed research across 11 domains, so you always know how strong the evidence is."
• A "Getting started" section:
  • "Have an existing course? /idstack course-import"
  • "Designing from scratch? /idstack needs-analysis"
  • "Run the full pipeline: /idstack pipeline"
• Then print the full command listing below

Always end with this command listing (after the welcome message):

Skills:
  needs-analysis         Three-level needs assessment (org, task, learner)
  learning-objectives    Measurable ILOs with Bloom's taxonomy + alignment check
  assessment-design      Rubrics, feedback strategies, formative checkpoints
  course-builder         Generate syllabus, modules, assignments, rubrics
  course-quality-review  Quality Matters audit + Community of Inquiry analysis
  accessibility-review   WCAG 2.1 AA + Universal Design for Learning review
  red-team               Adversarial audit across 5 dimensions
  course-export          Package for Canvas, Blackboard, Moodle (IMSCC/SCORM)
  course-import          Import from any LMS or authoring tool
  pipeline               Run all 8 skills in order (auto-skips completed)
  learn                  Search, promote, delete project learnings
  status                 Course health dashboard

More info: https://idstack.org

Stop after printing. Do not invoke any skill.

Important Rules

• Don't add your own workflow. Your only job is routing. Once you invoke a sub-skill, that skill takes over completely.
• Pass through args. If the user provided additional context after the subcommand name (e.g., /idstack learn search canvas), pass it as args to the sub-skill.
• Be forgiving with names. Accept common variations: "qa" → course-quality-review, "export" → course-export, "import" → course-import, "objectives" → learning-objectives, "build" → course-builder, "assess" → assessment-design.