Ed

Compatibility: AskUserQuestion falls back to numbered list on non-Claude-Code platforms.

ed

Author graduate-level, literature-grounded interactive textbooks as diffbook MDX — vetted sources, a dependency-ordered syllabus, scaffolded lessons, Bloom-distributed quizzes, and real-assignment projects, all behind scored quality gates.

Skill Requirements

command -v npx >/dev/null 2>&1 || echo "MISSING: npx"
# diffbook skill provides /diffbook init|author; used to scaffold and validate the book.

Inline Usage

Each verb operates on a {course} slug; planning artifacts live under .codevoyant/ed/{course}/, published MDX under the diffbook book (default book/).

/ed explore "transformer architectures"                 # vet & annotate reference materials
/ed plan-syllabus transformer-architectures             # dependency-ordered module program (gate ≥85)
/ed plan-module transformer-architectures 02            # lesson-level outline for one module (gate ≥80)
/ed create-lesson transformer-architectures 02 01       # author one lesson .mdx into the book
/ed create-quiz transformer-architectures 02            # author the module quiz .mdx
/ed create-project transformer-architectures 02         # author the module project + solution guide
/ed autodidact "reinforcement learning" --yes           # one-shot: brief → … → whole book, gated, best-effort
/ed update transformer-architectures 02/01              # re-run the minimal affected slice / apply annotations
/ed help

Bare-noun ergonomics: /ed syllabus …, /ed module …, /ed lesson …, /ed quiz …, /ed project … map to the plan-/create- verbs.

Critical Rules

• Never execute workflow logic here — this file only parses args and dispatches
• Step 0 always runs first — no exceptions
• Unknown verb → run help.md — never error silently
• Pass all remaining args through — the workflow receives $REMAINING_ARGS unchanged (course, module, lesson, and all --book/--dir/--yes flags)
• MDX output goes to the diffbook book, plan/draft artifacts go to .codevoyant/ed/{course}/ — see references/artifact-layout.md
• Markdown/MDX output: soft-wrap prose, never hard-wrap — every workflow and agent writes each paragraph as one continuous line; newlines separate paragraphs, list items, headings, and fences only. LaTeX uses \( \) / \[ \], never $…$. (Full guidance: references/pedagogy.md.)
• Model tiers, never model IDs — the ed-lesson-author agent declares metadata: model-tier: heavy; the platform maps tiers to concrete models (see references/model-tiers.md). Never hardcode a provider model ID (such as claude-*) in this skill.
• Gates before advancing — every stage scores its output and refuses to advance a weak artifact; interactive runs STOP on repeated failure, autodidact --yes logs to state.md and continues best-effort (see references/quality-gates.md)
• See references/workflows/ for per-verb behaviour; see references/ for shared references and templates

Step 0: Parse Arguments

The raw invocation args (filled by Claude Code / OpenCode slash commands): $ARGUMENTS. If this line is not filled in, read the verb and remaining args from the user's current message.

VERB="[first non-flag argument, or empty]"
REMAINING_ARGS="[everything after VERB, preserving order and flags]"

# Normalise bare-noun aliases to their full verbs
case "$VERB" in
  "")         VERB="help" ;;
  "syllabus") VERB="plan-syllabus" ;;   # /ed syllabus → /ed plan-syllabus
  "module")   VERB="plan-module" ;;     # /ed module   → /ed plan-module
  "lesson")   VERB="create-lesson" ;;   # /ed lesson   → /ed create-lesson
  "quiz")     VERB="create-quiz" ;;     # /ed quiz     → /ed create-quiz
  "project")  VERB="create-project" ;;  # /ed project  → /ed create-project
  "fix")      VERB="doctor" ;;          # /ed fix      → /ed doctor
  "repair")   VERB="doctor" ;;          # /ed repair   → /ed doctor
esac

Step 1: Dispatch to Workflow

Read and execute references/workflows/{VERB}.md, passing $REMAINING_ARGS as the argument string.

If references/workflows/{VERB}.md does not exist, fall back to references/workflows/help.md and note the unknown verb.

Workflow Index

• explore (references/workflows/explore.md) — find & vet reference materials for a course; verify URLs; annotate → explore/sources.md
• plan-syllabus (references/workflows/plan-syllabus.md) — dependency-ordered module program with Bloom objectives (gate ≥85) → syllabus.md
• plan-module (references/workflows/plan-module.md) — read a module's sources and craft a lesson-level outline (gate ≥80) → modules/{NN}/plan.md
• create-lesson (references/workflows/create-lesson.md) — author one lesson .mdx into the book, Feynman-style at graduate level
• create-quiz (references/workflows/create-quiz.md) — author a Bloom-distributed module quiz .mdx
• create-project (references/workflows/create-project.md) — author a module project + solution guide grounded in a real sourced assignment
• autodidact (references/workflows/autodidact.md) — chain everything from a topic into a whole book, gated, with a state.md ledger
• update (references/workflows/update.md) — re-run the minimal affected slice; consume <!-- > --> / <!-- >> --> annotations
• doctor (references/workflows/doctor.md) — repair a mis-scaffolded book in place (project at root, book/ as contentPath); dry-run by default, --fix to apply
• help (references/workflows/help.md) — print command reference

Agent Index

• ed-lesson-author (agents/ed-lesson-author.md) — authors ONE lesson .mdx from {module}/plan.md + the module's source shortlist; self-checks against the lesson gate; used by create-lesson (single) and autodidact (per-lesson fan-out)