| name | pipeline-status |
| description | Report NotebookLM video pipeline progress as a formatted ✅/⚠️ checklist. Trigger: user asks 'status', 'check progress', 'what's running', or 'which pipeline is running'. |
Pipeline Status Report
Report progress of all active notebooklm-pipeline runs as a formatted checklist.
Workflow Steps
When this skill is invoked, follow these steps in order:
Step 1: Query Live State
For each active pipeline, check the current state (no waiting, just a snapshot):
nlm studio status <notebook-id> -j
nlm research status <notebook-id>
Also inspect local artifacts (download/watermark/trim/meta/cover/prepare) to know which post-video steps are done.
Step 2: Report Progress Checklist
Output a formatted list of all pipeline steps. Every pipeline run must report progress in this exact shape:
Notebook #{Index}: {Title}
✅ 1. Research <notebook title>
✅ 2. Video generation
✅ 3. Download + watermark + trim
⚠️ 4. Meta files (title/desc/tags)
⚠️ 5. Cover image
⚠️ 6. Prepare upload
⚠️ 7. Publish
Rules:
- One line per pipeline step, numbered in pipeline order (1-7).
- ✅ = step completed.
- ⚠️ = step pending / in progress (brief note after the line).
- No decorative headers, no raw JSON dumps — just the checklist.
- Include live artifact/notebook IDs on the completed step's line only when useful.
- Multiple pipelines: one block per pipeline, labeled
Notebook #{Index}: {Title}.
Step 3: Surface What Needs Attention
After the checklist, one line stating the blocking next action per pipeline (e.g. "P2 next: import sources"). Ask user which to continue if choices exist.
Example
User: "check status"
Agent:
- Queries
nlm studio status for each active notebook.
- Checks which local files exist (video.mp4, title.json, cover.png, etc.).
- Reports:
Notebook #1: 同济大学取消教师长期聘任
✅ 1. Research
✅ 2. Video generation
✅ 3. Download + watermark + trim
⚠️ 4. Meta files — pending
⚠️ 5. Cover image — pending
⚠️ 6. Prepare upload — pending
⚠️ 7. Publish — pending
