hxddh / object storage workbench
S3 automation toolkit
Tools for object storage work that should be fast, inspectable, and easy to automate: inventory large buckets, diff object state, give Pi a diagnosis layer, and inspect HTTP traffic when the SDK hides the clue.
Storage operations flow
Storage workflow
Use s3-turbo-list when inventory or diff should be automatic, resumable, and machine-readable. Use StorageOps when Pi should diagnose an S3-compatible failure with skills and tools. Use httpmon when the important clue is below the SDK, inside real HTTP requests and responses.
s3-turbo-list
Run automatic inventory or diff on large S3-compatible buckets, then keep Parquet, JSON plans, and run manifests.
StorageOps
Give Pi a storage diagnosis layer: domain skills, bounded tools, deterministic helpers, and eval-backed routing.
httpmon
Capture the real HTTP request and response when signing, retries, redirects, or provider behavior hide the clue.
Product evidence
Evidence before features
The main products lead with checkable evidence: one shows benchmark data, the other shows what the agent package installs.
1M objects, ~55s to ~19s
The current README cites a third-party Alibaba Cloud OSS benchmark: 1M objects in ~55s single-stream and ~19s with `-c 8`.
The storage diagnosis layer for Pi
StorageOps is not a prompt pack. It gives Pi skills, tools, helpers, and evals that constrain the diagnostic path and evidence boundary.
Storage operations
List, diff, diagnose
Choose this when large-bucket inventory or diff should run automatically, leave Parquet artifacts, and expose JSON plans or manifests for CI and agents.
StorageOpsChoose this when Pi should diagnose S3-compatible failures using object-storage skills, extension tools, helper scripts, and eval-backed routing.
https-traffic-inspectorChoose this when the important clue is inside the HTTP request or response hidden by curl, aws, Python, Node, or another CLI.
Browser utilities
Small browser tools
These two tools stay intentionally narrow: one checks time, the other adds rating context where you are already reading.
