get-full-content

Files (2)

• SKILL.md
• get-full-content.js

SKILL.md (first lines)

# Get Full Content

This is an **Action** skill — a single, unambiguous API call. It is deterministic and narrowly scoped.

## Endpoints

- **Multi-section courses / learning paths:** `GET /incoming/v2/fullContent/{type}/{id}` where `{type}` is `courses` or `learningPaths`.
- **Single-page `article` / `video` course kinds:** `GET /incoming/v2/content/courses/{id}` — returns `{ id, title, slug, body }` (the article HTML, or the video embed iframe/script markup). These kinds store their HTML on the course itself; `fullContent` does **not** return the video body and buries the article body, so this endpoint is used for them.

This action picks the right endpoint for you (see [Single-page content](#single-page-article--video-content)).

## When to use

- **Reading the actual HTML body of existing topics** — "show me the content of lesson X", "what does this course say", before rewriting existing content.
- Reading the body of a single-page **article** or **video** course.
- Exporting/auditing a whole course's content in one call.

> **Important:** `get-topic` and `get-topics-by-lesson` return topic *metadata* (id, title, type, pre/post text) but **do not return the `body` HTML**. To read the actual lesson/topic content of existing material you **must** use this action. This is the endpoint that fixes "ask Claude to show us lesson content and it fails."

## Agent instructions

1. Require the `id`:
   - For a **course**, this is the **course ID within a course group** (a.k.a. `displayCourseId`), **not** the course group ID. Get it from `list-course-groups`/`get-course-group` (`displayCourseId`) or via the `resolve-content` action. It is the same ID `get-course-structure` uses.
   - For a **learning path**, this is the learning path ID.
2. Default `--type=courses`. Pass `--type=learningPaths` for a learning path.
3. Run `get-full-content.js --id=UUID`. You do **not** need to know whether the course is a normal multi-section course or a single-page article/video — it is auto-detected.
4. Return the topic bodies. Optionally pass `--lesson=LESSON_UUID` to focus the derived tree on a single lesson.

## Run the script

```bash
node "${CLAUDE_PLUGIN_ROOT:+$CLAUDE_PLUGIN_ROOT/skills/actions/get-full-content/}get-full-content.js" --id=COURSE_UUID
node "${CLAUDE_PLUGIN_ROOT:+$CLAUDE_PLUGIN_ROOT/skills/actions/get-full-content/}get-full-content.js" --id=COURSE_UUID --lesson=LESSON_UUID
node "${CLAUDE_PLUGIN_ROOT:+$CLAUDE_PLUGIN_ROOT/skills/actions/get-full-content/}get-full-content.js" --id=PATH_UUID --type=learningPaths
# Force the single-page path (usually unnecessary — it is auto-detected):
node "${CLAUDE_PLUGIN_ROOT:+$CLAUDE_PLUGIN_ROOT/skills/actions/get-full-content/}get-full-content.js" --id=COURSE_UUID --kind=article
node "${CLAUDE_PLUGIN_ROOT:+$CLAUDE_PLUGIN_ROOT/skills/actions/get-full-content/}get-full-content.js" --id=COURSE_UUID --single-page
```

## Single-page article / video content

A course group whose `kind` is `article` or `video` is **single-page**: its HTML/embed body lives on the course, not in a section/lesson/topic tree. This action handles that transparently:

- The normal `fullContent` path is tried first. When the response looks single-page (exactly one section → one lesson → at most one topic), the script confirms and reads the real body via `GET /content/courses/{id}` and prints it as a `_singlePage` block:

```json
_singlePage (article/video body via GET /content/courses/{id}):
{
  "id": "COURSE_UUID",
  "title": "Cloud Cost Optimization Strategies for AWS",
  "slug": "cloud-cost-optimization-strategies-for-aws",
  "body": "<h2>Controlling Cloud Spend</h2><p>…</p>"
}
```

- **Override:** pass `--kind=article`, `--kind=video`, or `--single-page` to read `GET /content/courses/{id}` directly (skipping `fullContent`). Useful when you already know the kind (e.g. from `resolve-content`, which returns it).
- `GET /content/courses/{id}` returns `{ errors: [...], data: null }` (HTTP 200) for content that is **not** single-page, so the override cleanly falls back to the multi-section path when misapplied.

## Response shape

The raw API response is a **flattened, dotted-key object**, e.g.:

```json
{
  "id": "COURSE_UUID",
  "title": "My Course",
  "section.0.id": "SECTION_UUID",
  "section.0.title": "Section 1",
  "section.0.lesson.0.id": "LESSON_UUID",
  "section.0.lesson.0.title": "Lesson 1",
  "section.0.lesson.0.topic.0.id": "TOPIC_UUID",
  "section.0.lesson.0.topic.0.title": "Intro",
  "section.0.lesson.0.topic.0.body": "<p>Welcome…</p>"
}
```

For courses, this script also prints a `_tree:` view that nests the flattened keys into `sections → lessons → topics` so the content is easy to scan and to line up topic `id`s with their `body` for a follow-up `update-course-content` call.

## Limits

- Only `courses` and `learningPaths` are supported. Any other `type` returns HTTP 400.
- This endpoint is **rate limited** (`X-RateLimit-*` headers; HTTP 429 when exceeded). Prefer one full-content read over many per-topic reads.

<HARD-GATE>
Do not modify, create, or delete any content. This is a read-only action.
</HARD-GATE>

From thoughtindustries/ti-admin-plugin, MIT.

More from thoughtindustries/ti-admin-plugin