Tencent Cloud COS and CI Manager
Paste the following prompt into your AI chat to install this skill:
Please follow https://skillhub.cn/install/skillhub.md to install @org-m6xe1azv/skill-test2.
About this skill
Problem Solved
When object storage, image processing, document conversion, content moderation, speech processing, and MetaInsight search must enter the same automation flow, calling cos-nodejs-sdk-v5 piecemeal can make scripts hard to maintain: credentials are scattered, async task polling is inconsistent, dataset templates are easy to mismatch, and outputs are not directly consumable.
How the Skill Works
The skill wraps scripts/cos_node.mjs into stable action calls with unified parameters and JSON output. Core capabilities include:
- COS operations: upload, download, list, sign-url, delete-multiple, copy-object;
- Bucket management: list/create buckets, configure ACL, CORS, tags, versioning, and lifecycle, while deleting or emptying buckets is forbidden;
- CI processing: image processing, OCR, document-to-PDF, video transcoding, smart cover, content moderation, speech recognition/synthesis, compression, and decompression;
- MetaInsight: manage datasets by template, support image-to-image search, text-to-image search, face search, metadata search, and multimodal document search;
- Knowledge-base flow: create a knowledge base, upload documents, and run semantic search.
On first use, it guides credential setup, prefers STS temporary credentials, and can encrypt .env into .env.enc. Async tasks are polled by default, and environment checks can isolate setup issues.
Boundaries
It fits automation pipelines or internal tooling for Tencent Cloud COS/CI. It is not a general cloud resource manager: it does not delete or empty buckets, different search types require matching dataset templates, and permissions are limited by the child-account policy. For one-off API debugging, the official SDK may be simpler; for composable, auditable multi-step object-storage workflows, this skill is more suitable.
Use Cases
- Batch-upload images before release, generate signed URLs, and audit violating content
- Convert PDFs and Word files to PDF or preview images for internal document search
- Run image-to-image, text-to-image, or metadata search with the correct dataset template
- Create a knowledge base, upload documents, and return semantic search summaries with sources
Best For
- Backend engineers maintaining object-storage pipelines who need upload, signing, and moderation as JSON APIs
- Algorithm engineers building enterprise document search who need MetaInsight datasets and image-to-image search
- Product engineers on content-platform risk teams who need violation detection for images, videos, and text
- Full-stack engineers building internal knowledge bases who need to create bases, upload docs, and run semantic search
Related Skills
Automatically indexes Gradle-cached AAR/JAR dependency classes and returns library coordinates, versions, and public APIs by fully qualified name, using only the Python standard library.
Codifies AMT and YourMT3 training conventions, script patterns, hyperparameters, precision, checkpoints, and NaN safeguards.
Retrieve relevant chunks from a customer-managed PKM dataset by dataset_id and return concise, source-annotated answers.
Convert PRDs, user stories, or functional specs into prioritized test-point checklists covering functional, business-rule, boundary, exception, and non-functional dimensions.