ai notes of video

Generate AI-powered notes from videos (document, outline, or graphic-text formats)

Security Scan

The skill's code, required environment variable (BAIDU_API_KEY), and runtime behavior align with its description: it sends video URLs to Baidu's AI note service and polls for results — nothing in the package attempts to access unrelated credentials or system areas.

Purpose & Capability

The skill's name/description match the included scripts and required env var. It calls Baidu endpoints (qianfan.baidubce.com) and requires BAIDU_API_KEY, which is appropriate for a Baidu AI service integration.

Instruction Scope

SKILL.md and the scripts limit actions to submitting a public video URL, polling task status, and printing results. The instructions do not request or read unrelated files, credentials, or system state beyond BAIDU_API_KEY and command arguments. Network calls are to the documented Baidu endpoints only.

Install Mechanism

There is no install spec (instruction-only), which minimizes install risk. However, the included Python scripts import the 'requests' library but the skill does not declare dependencies or an install step to ensure 'requests' is present — users must ensure python3 + requests are available in the runtime.

Credentials

Only BAIDU_API_KEY is required and used consistently as a bearer token for the Baidu API. No unrelated secrets or environment paths are requested.

Persistence & Privilege

The skill does not request always:true and does not modify other skills or system-wide settings. It runs only when invoked and uses only the declared env var.

Assessment

This skill appears to do what it says: it sends a public video URL to Baidu's 'ai_note' service and polls for results. Before installing, consider: 1) You will need a BAIDU_API_KEY — do not reuse highly privileged credentials; restrict the key if possible. 2) The scripts make outbound network requests to qianfan.baidubce.com and will transmit video URLs (and thus the referenced content) to Baidu — do not send sensitive/private videos unless you trust the service and understand its data retention/privacy policy. 3) Ensure the runtime has python3 and the 'requests' library installed (the package does not declare dependencies). 4) The code has minor inconsistencies in error-field handling and messages (implementation bugs, not malicious). If you are comfortable with these privacy/network implications and can provide a dedicated BAIDU_API_KEY, this skill is coherent with its stated purpose.

AI Video Notes

Generate structured notes from video URLs using Baidu AI. Supports three note formats.

Workflow

1. Create Task: Submit video URL → get task ID
2. Poll Status: Query task every 3-5 seconds until completion
3. Get Results: Retrieve generated notes when status = 10002

Status Codes

Code	Status	Action
10000	In Progress	Continue polling
10002	Completed	Return results
Other	Failed	Show error

Note Types

Type	Description
1	Document notes
2	Outline notes
3	Graphic-text notes

APIs

Create Task

Endpoint: POST /v2/tools/ai_note/task_create

Parameters:

• video_url (required): Public video URL

Example:

python3 scripts/ai_notes_task_create.py 'https://example.com/video.mp4'

Response:

{
  "task_id": "uuid-string"
}

Query Task

Endpoint: GET /v2/tools/ai_note/query

Parameters:

• task_id (required): Task ID from create endpoint

Example:

python3 scripts/ai_notes_task_query.py "task-id-here"

Response (Completed):

{
  "status": 10002,
  "notes": [
    {
      "tpl_no": "1",
      "contents: ["Note content..."]
    }
  ]
}

Polling Strategy

Option 1: Manual Polling

1. Create task → store task_id
2. Query every 3-5 seconds:

   python3 scripts/ai_notes_task_query.py <task_id>
   

3. Show progress updates:
   • Status 10000: Processing...
   • Status 10002: Completed
4. Stop after 30-60 seconds (video length dependent)

Option 2: Auto Polling (Recommended)

Use the polling script for automatic status updates:

python3 scripts/ai_notes_poll.py <task_id> [max_attempts] [interval_seconds]

Examples:

# Default: 20 attempts, 3-second intervals
python3 scripts/ai_notes_poll.py "task-id-here"

# Custom: 30 attempts, 5-second intervals
python3 scripts/ai_notes_poll.py "task-id-here" 30 5

Output:

• Shows real-time progress: [1/20] Processing... 25%
• Auto-stops when complete
• Returns formatted notes with type labels

Error Handling

• Invalid URL: "Video URL not accessible"
• Processing error: "Failed to parse video"
• Timeout: "Video too long, try again later"