@studiolxd/lxd-cli 0.1.0-next.0

package/LICENSE

@@ -0,0 +1,36 @@
+Copyright (c) 2026 Studio LXD. All rights reserved.
+
+PROPRIETARY SOFTWARE — NOT OPEN SOURCE
+
+This software and its source code (the "Software"), distributed as the npm
+package "@studiolxd/lxd-cli", are the proprietary property of Studio LXD.
+Studio LXD retains all right, title, and interest in and to the Software.
+
+The Software may be made publicly available on a package registry (e.g., npm)
+for distribution convenience. Public availability does NOT constitute an
+open-source license and does NOT grant any rights beyond those expressly stated
+below or otherwise expressly granted in writing by Studio LXD.
+
+Downloading, installing, or accessing the Software from npm or any other source
+does NOT grant you any right to:
+
+  - copy, reproduce, or store the Software except as strictly required to run it
+    for its intended purpose;
+  - modify, adapt, translate, or create derivative works of the Software;
+  - redistribute, republish, sublicense, sell, rent, lease, or otherwise
+    transfer the Software or access to it;
+  - remove or alter any proprietary notices.
+
+You may use the Software only under, and subject to, the terms provided by
+Studio LXD. Any use not expressly permitted is prohibited.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
+FOR A PARTICULAR PURPOSE, AND NONINFRINGEMENT. IN NO EVENT SHALL STUDIO LXD BE
+LIABLE FOR ANY CLAIM, DAMAGES, OR OTHER LIABILITY ARISING FROM, OUT OF, OR IN
+CONNECTION WITH THE SOFTWARE OR ITS USE.
+
+Studio LXD may change these license terms at any time, including adopting an
+open-source or open-core model for future versions, at its sole discretion.
+
+For licensing inquiries, contact Studio LXD.

package/README.md

@@ -0,0 +1,123 @@
+# @studiolxd/lxd-cli
+
+A **local-first, CLI-first** authoring tool that takes you from a *learning goal* to a
+*publishable elearning package* — built by **your own external AI coding agent** (Claude Code,
+Cursor, Codex, …). The CLI doesn't replace the agent; it **orchestrates** the workflow: capture a
+structured package specification, generate agent-ready handoff materials, preview locally, validate,
+and export to multiple targets.
+
+> Not a SaaS, not a visual editor, not a block builder, not tied to one AI provider, one frontend
+> framework, or one LMS standard.
+
+## Status
+
+Implements the full v1 vertical: US1 (single-experience → web), US2 (SCORM tracking + export),
+US3 (multi-experience), US4 (extensible official mechanics), US5 (advanced design direction).
+
+## Requirements
+
+- Node.js ≥ 20 (macOS / Linux / Windows)
+- Your own AI coding agent for the implementation step (run manually)
+
+## Install
+
+```bash
+npm install        # install deps
+npm run build      # bundle the CLI to dist/ (+ copy runtime assets)
+npm link           # optional: expose `lxd` globally for local dev
+# or run from source during dev:
+npx tsx src/cli/index.ts <command>
+```
+
+## Quickest start — the guided wizard
+
+```bash
+lxd create my-course     # (alias: lxd wizard) one guided, interactive session:
+                               # answer prompts → buffers in memory → writes the package in one
+                               # commit → optional validation (default yes) + handoff (default no)
+```
+Create-new only and non-destructive: on an existing package it asks before replacing; cancelling
+before the commit leaves zero files. The wizard reuses the same core as the commands below.
+
+## The workflow (individual commands)
+
+```bash
+lxd init my-course && cd my-course   # scaffold a package project
+lxd goal       -g "Recognize workplace hazards"
+lxd audience   -d "New staff" -m 20
+lxd experience add --id spot --purpose "Find hazards" --scenario "aisle" --mechanic risk-detection
+lxd structure  --sync
+lxd design     -l brief -b "Clean, accessible, corporate"
+lxd evaluation -c all-required
+lxd tracking   --report completion score success
+lxd validate                          # actionable pass/fail across all checks
+lxd handoff                           # write .agent/** specs + prompts + scaffold generated/
+#  → run YOUR agent against .agent/** to produce generated/
+lxd preview --mock-lms                # build + serve locally, with a mock LMS
+lxd export --target web --target scorm-1.2 --target scorm-2004
+```
+
+See **[docs/commands.md](docs/commands.md)** for the full command reference and
+**[specs/001-elearning-package-cli/quickstart.md](specs/001-elearning-package-cli/quickstart.md)**
+for runnable scenarios.
+
+## Concepts
+
+- **Package** is the top-level unit; it contains **one or more experiences**.
+- **Experience mechanics** are official, declarative manifests (`src/mechanics/manifests/`), with an
+  optional code plugin for advanced cases. New official mechanics are added without core changes.
+- **Design direction** is a first-class, multi-level input (brief → tokens → patterns →
+  design-system → full constraints), carried into the agent handoff.
+- **Tracking** uses per-experience signals aggregated into a single LMS state via explicit roll-up
+  rules.
+- **Adapters** are the extension seams, behind documented contracts:
+  - Framework adapter (v1: Vanilla JS + Web Components)
+  - Export adapters (v1: `web`, `scorm-1.2`, `scorm-2004`)
+- **Agent handoff** is the reserved boundary: v1 produces materials for manual use; programmatic
+  invocation is a future extension. The handoff transmits the **full instructional intent** —
+  audience, package tone, and per-experience purpose, scenario, mechanic guidance (incl. plugin
+  enrichment), feedback, and tracking — and instructs the agent to treat the materials as the
+  source of truth (not to infer the learning design from code).
+
+## External assets (not vendored)
+
+- [`@studiolxd/scorm`](https://github.com/studiolxd/scorm) — SCORM runtime/tracking, consumed by
+  generated projects as a **versioned dependency**. SCORM-targeted packages **must** use it: the
+  handoff forbids hand-rolling a SCORM runtime, and `validate` **fails** if `generated/` contains a
+  custom runtime (`window.API`/`LMS*`/2004 handlers) instead of `@studiolxd/scorm`.
+- [`scorm-skills`](https://github.com/studiolxd/scorm-skills) — canonical SCORM guidance, referenced
+  as a **versioned snapshot** in `.agent/scorm-skills/` when a SCORM target is selected.
+
+Both are external and versioned; the CLI never vendors or forks them. Generated projects update
+their versions only via `lxd assets update` — never silently.
+
+## Develop
+
+```bash
+npm run typecheck   # tsc --noEmit
+npm run lint        # eslint
+npm test            # vitest (unit + contract + integration)
+npm run build       # esbuild bundle + asset copy
+```
+
+CI runs typecheck/lint/build/test on macOS, Linux, and Windows (Node 20).
+
+## Release (manual)
+
+The package publishes to npm as **`@studiolxd/lxd-cli`** under the **`next`** dist-tag (never
+`latest`). It is **proprietary** (`UNLICENSED`, all rights reserved) — public on npm ≠ open source.
+Nothing here publishes automatically; publishing is a deliberate manual step.
+
+```bash
+npm run verify:pack        # pack → install the tarball in a clean temp dir → run the installed lxd
+                           # (--help/--version/mechanic list/init). Needs network for dependency install.
+npm publish --tag next --access public   # manual; run only after verify:pack passes
+```
+
+`prepublishOnly` re-runs typecheck + lint + tests before any publish, and `prepack` rebuilds `dist/`,
+so a broken or stale package cannot be published. After publishing, real-world use:
+
+```bash
+npx @studiolxd/lxd-cli@next --help
+npm install -g @studiolxd/lxd-cli@next && lxd create
+```