AgentSkillsCN

s3

管理Amazon S3、Google Cloud Storage或Azure Blob Storage中的文件。列出、上传、下载并预览云存储桶中的对象。

SKILL.md
--- frontmatter
name: s3
description: >
  Manage files in Amazon S3, Google Cloud Storage, or Azure Blob Storage.
  List, upload, download, and preview objects from cloud storage buckets.
metadata:
  openclaw:
    requires:
      bins: [clawdata]
    primaryEnv: AWS_DEFAULT_REGION
    tags: [cloud, storage, s3, gcs, azure, blob, files]

S3 / GCS / Azure Blob

Cloud storage skill for managing objects across AWS S3, Google Cloud Storage, and Azure Blob Storage.

Commands

TaskCommand
List bucketsclawdata s3 ls
List objectsclawdata s3 ls s3://bucket/prefix/
Upload fileclawdata s3 cp local.csv s3://bucket/data/
Download fileclawdata s3 cp s3://bucket/data/file.csv ./local/
Preview fileclawdata s3 head s3://bucket/data/file.csv
Remove objectclawdata s3 rm s3://bucket/data/file.csv
Sync directoryclawdata s3 sync ./data/ s3://bucket/data/

Supported Providers

ProviderURI SchemeAuth Env Vars
AWS S3s3://AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY
Google GCSgs://GOOGLE_APPLICATION_CREDENTIALS
Azure Blobaz://AZURE_STORAGE_CONNECTION_STRING

When to use

  • •User wants to load data from a cloud bucket → clawdata s3 cp then clawdata data ingest
  • •User wants to export query results to S3 → clawdata db export then clawdata s3 cp
  • •User asks about files in a bucket → clawdata s3 ls

DuckDB Integration

DuckDB's httpfs extension supports S3 natively:

sql
-- Load directly from S3
SELECT * FROM read_parquet('s3://my-bucket/data/*.parquet');

Use clawdata db query with S3 paths when httpfs is loaded.