alicloud-skills

Alibaba Cloud Core AI Agent Skills

Language

English (current) | 简体中文 | 繁體中文

Quick links: Quick Start | Skill Index

Cloud Mind: fold world-class cloud infrastructure into your AI chat box. AI云：把世界级云基础设施带进你的 AI 对话框。

A curated collection of Alibaba Cloud core AI Agent skills covering key product lines, including Model Studio, OSS, ECS, and more.

Latest Coverage

• DashScope video generation now includes Wan, Kling v3, Vidu, Animate-Move, and video style repaint workflows.
• Network operations now include a dedicated aliyun-vpc-manage skill for VPC/VSwitch inventory and lifecycle tasks.
• Root README skill indexes are generated from skills/**/SKILL.md so new skills stay synchronized across English, Simplified Chinese, and Traditional Chinese docs.

Quick Start

Recommended install (all skills, skip prompts, overwrite existing):

npx skills add cinience/alicloud-skills --all -y --force

If you still see a selection prompt, press a to select all, then press Enter to submit.

Use a RAM user/role with least privilege. Avoid embedding AKs in code or CLI arguments.

Configure AccessKey (recommended):

export ALIBABACLOUD_ACCESS_KEY_ID="your-ak"
export ALIBABACLOUD_ACCESS_KEY_SECRET="your-sk"
export ALIBABACLOUD_SECURITY_TOKEN="your-sts-token" # optional, for STS
export ALIBABACLOUD_REGION_ID="cn-beijing"
export DASHSCOPE_API_KEY="your-dashscope-api-key"

Environment variables take precedence. If they are not set, the CLI/SDK falls back to ~/.alibabacloud/credentials. ALIBABACLOUD_REGION_ID is an optional default region; if unset, choose the most reasonable region at execution time, and ask the user when ambiguous.

Default recommendation is the ALIBABACLOUD_* prefix. Runtime scripts in this repo also accept legacy ALIBABA_CLOUD_* and ALICLOUD_* aliases for compatibility.

If env vars are not set, use standard CLI/SDK config files:

~/.alibabacloud/credentials

[default]
type = access_key
access_key_id = your-ak
access_key_secret = your-sk
dashscope_api_key = your-dashscope-api-key

For STS, set type = sts and add security_token = your-sts-token.

Examples (Docs Review & Benchmark)

1. Product docs + API docs review

• Prompt: "Use aliyun-platform-docs-review to review product docs and API docs for Bailian, then return prioritized P0/P1/P2 improvements with evidence links."

2. Multi-cloud comparable benchmark

• Prompt: "Use aliyun-platform-docs-benchmark to benchmark Bailian against Alibaba Cloud/AWS/Azure/GCP/Tencent Cloud/Volcano Engine/Huawei Cloud with preset llm-platform, and output a score table plus gap actions."

Standalone Skills With Prompt Examples

1. Text-to-image (Qwen Image)

• Demo: Generate an image
• Prompt: "Use aliyun-qwen-image to generate a 1024*1024 poster with a minimalist coffee theme, output filename poster.png."

2. Image-to-video (Wan Video)

• Demo: Generate a 4-second video from a reference image (public image URL required)
• Prompt: "Use aliyun-wan-video with reference image https://.../scene.png to generate a 4-second, 24fps, 1280*720 shot. Prompt: morning city timelapse."

3. VPC inventory and provisioning

• Demo: List VPCs in a region or scaffold a new VPC/VSwitch flow
• Prompt: "Use aliyun-vpc-manage to list all VPCs and VSwitches in cn-hangzhou, then suggest a /16 VPC CIDR and /24 VSwitch layout for a two-AZ production deployment."

4. Kling V3 video generation

• Demo: Create a storyboard or image-to-video task with DashScope Kling
• Prompt: "Use aliyun-kling-video to generate a 5-second 16:9 clip from prompt a neon-lit alley in the rain, and explain when to choose the omni model instead of the standard model."

5. Vidu video generation

• Demo: Compare text-to-video vs keyframe-to-video request shapes
• Prompt: "Use aliyun-vidu-video to show the minimum request bodies for vidu/viduq3-pro_text2video and vidu/viduq3-pro_start-end2video, including recommended resolution, size, and duration."

6. Text-to-speech (Qwen TTS)

• Demo: Generate speech audio with DashScope
• Prompt: "Use aliyun-qwen-tts to synthesize this paragraph, voice=Cherry, language=English, and return the audio URL."

7. Document structure parsing (DocMind)

• Demo: Parse title/paragraph structure from PDF
• Prompt: "Use aliyun-docmind-extract to parse this PDF (URL: ...), and return structured results."

8. Vector retrieval (DashVector)

• Demo: Create collection, insert, and query
• Prompt: "Use aliyun-dashvector-search to create a collection with dimension=768, insert 2 records, then run a topk=5 query."

9. OSS upload/sync (ossutil)

• Demo: Upload a local file to OSS
• Prompt: "Use aliyun-oss-ossutil to upload ./local.txt to oss://xxx/path/."

10. SLS log troubleshooting

• Demo: Query 500 errors in the last 15 minutes
• Prompt: "Use aliyun-sls-log-query to find 500 errors in the last 15 minutes and aggregate by status."

11. FC 3.0 quick deployment (Serverless Devs)

• Demo: Initialize and deploy a Python function
• Prompt: "Use aliyun-fc-serverless-devs to initialize an FC 3.0 Python project and deploy it."

12. Content safety moderation (Green)

• Demo: Discover/call moderation APIs via OpenAPI
• Prompt: "Use aliyun-green-moderation to list available APIs first, then provide a minimal text-moderation parameter example."

13. KMS key management

• Demo: List keys or create a key
• Prompt: "Use aliyun-kms-manage to provide an OpenAPI parameter template for creating a symmetric key."

14. Automated product docs and API docs review

• Demo: Auto-fetch latest docs by product and output improvement suggestions
• Prompt: "Use aliyun-platform-docs-review to review product docs and API docs for Bailian, and output prioritized P0/P1/P2 suggestions with evidence links."

15. Cross-cloud docs/API benchmark

• Demo: Compare similar products across Alibaba Cloud/AWS/Azure/GCP/Tencent Cloud/Volcano Engine/Huawei Cloud
• Prompt: "Use aliyun-platform-docs-benchmark for Bailian with preset llm-platform, and output a score table and gap recommendations."

Solution Playbooks (Scenario Prompt Templates)

1. Marketing asset pipeline (image -> video -> voice-over -> upload)

Template: "Chain these skills in order:

1. aliyun-qwen-image to generate a poster image (theme: {theme}, size: {size}).
2. aliyun-wan-video to generate a {duration}s video from that image (fps={fps}, size={size}, shot description: {shot_desc}).
3. aliyun-qwen-tts to synthesize narration (voice={voice}, text: {narration_text}, language={language}).
4. aliyun-oss-ossutil to upload video and audio to {oss_path}. Return final asset URLs and descriptions."

2. Customer-support knowledge retrieval + voice answer

Template: "Use aliyun-docmind-extract to parse document (URL: {doc_url}) into structured content; then aliyun-dashvector-search to build and ingest the vector index; then answer user question {user_question} with topk={topk} retrieval and generate spoken response using aliyun-qwen-tts (voice={voice}, language={language}). Return text answer + audio URL."

3. Content moderation + publishing

Template: "Use aliyun-green-moderation to moderate content (type: {text|image|video}, content: {content_or_url}). If passed, upload with aliyun-oss-ossutil to {oss_path} and return public URL. If failed, return reason and replacement suggestions."

4. Site log troubleshooting + automated alert

Template: "Use aliyun-sls-log-query to query errors in {time_range} (query: {query}), aggregate by {aggregate_field}, and determine if threshold {threshold} is exceeded. If exceeded, use aliyun-fc-serverless-devs to trigger alert function (function: {function_name}, params: {alert_params}). Return stats and alert status."

5. Multilingual content production (generate -> translate -> voice-over)

Template: "Use aliyun-aicontent-generate to generate copy (topic: {topic}, style: {style}, length: {length}); use aliyun-anytrans-translate to translate to {target_language}; use aliyun-qwen-tts for speech (voice={voice}, language={language}). Return original text, translated text, and audio URL."

6. Training data cleaning and archiving

Template: "Run compliance check with aliyun-green-moderation (content: {content_or_url}). If passed, run structured extraction with aliyun-docmind-extract when applicable. Finally archive to {oss_path} with aliyun-oss-ossutil, and return archive manifest + URLs."

7. Log metric analysis report

Template: "Use aliyun-sls-log-query in {time_range} with {query|analysis}, output a stats table by {dimension}, then summarize a visualization report using aliyun-gbi-analytics (metrics: {metrics}, dimensions: {dimensions}). Return key metrics and report summary."

8. Business search and recommendation

Template: "Use aliyun-dashvector-search for intent-based vector retrieval (topk={topk}, filter={filter}), then aliyun-airec-manage for reranking and supplementary recommendations (strategy: {strategy}). Return final recommendation list and rationale."

9. Enterprise call scenario (call center + chatbot + voice)

Template: "Use aliyun-ccc-manage to create/route inbound calls (number: {number}, routing: {routing_strategy}); use aliyun-chatbot-manage for FAQ hit or handoff decision; use aliyun-qwen-tts to play responses (voice={voice}, language={language}). Return final script and audio URL."

10. Security and compliance closed loop (key management + audit)

Template: "Use aliyun-kms-manage to create/manage keys (purpose: {purpose}, alias: {alias}); use aliyun-sls-log-query to query security audit logs in {time_range} (query: {query}); if anomalies are found, return mitigation suggestions or trigger an alert function ({function_name}). Return key status, audit results, and remediation suggestions."

Repository Structure

• skills/ — canonical skill sources grouped by product line
  • ai/ — Model Studio (capability-based groups)
    • text/ image/ audio/ video/ multimodal/ search/ recommendation/ content/ service/ translation/ platform/ misc/ entry/
  • backup/ — BDRC / HBR
  • compute/ — ECS / FC / SWAS
  • data-analytics/ — DataAnalysisGBI
  • data-lake/ — DLF
  • database/ — AnalyticDB / RDS
  • media/ — intelligent media creation
  • network/ — CDN / DNS / VPC / ALB / ESA
  • observability/ — SLS
  • platform/ — CLI / OpenAPI / docs workflows
  • security/ — content moderation / firewall / host security / identity / key management
  • storage/ — OSS
• examples/ — end-to-end stories and usage walkthroughs

Brand Aliases

• modelstudio/ — symlink to skills/ai/ (overseas brand)

Governance Checks

Run repository governance checks locally:

bash scripts/verify_governance.sh

Workflow gate (prepare -> validate -> deploy):

bash scripts/workflow_prepare.sh demo-20260303
bash scripts/workflow_validate.sh demo-20260303
bash scripts/workflow_deploy.sh demo-20260303

Standards and migration notes:

• docs/standards/skill-template.md
• docs/standards/skill-template-migration.md
• docs/standards/coverage-policy.md
• docs/standards/workflow-gate.md

Skill Index

Category	Skill	Description	Path
ai/audio	aliyun-cosyvoice-voice-clone	Use when creating cloned voices with Alibaba Cloud Model Studio CosyVoice customization models, especially cosyvoice-v3.5-plus or cosyvoice-v3.5-flash, from reference audio and then reusing the returned voice_id in later TTS calls.	skills/ai/audio/aliyun-cosyvoice-voice-clone