shelby_skill

This skill enables decentralized blob storage on Aptos with erasure coding, SDKs, and private network access for scalable video, AI, and datasets.

62

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill raintree-technology/claude-starter --skill shelby

  • SKILL.md5.0 KB

Overview

This skill is a Shelby Protocol expert for decentralized blob storage on the Aptos blockchain. It explains erasure coding with Clay Codes, the TypeScript SDK and clients, smart contract interactions, CLI workflows, storage provider topology, and dApp integration patterns. The focus is on building high-performance storage solutions for video streaming, AI training, and large datasets.

How this skill works

The skill inspects the Shelby architecture: RPC servers that handle blob operations, a private fiber network connecting provider servers, and Aptos smart contracts for state, audits, and settlements. It explains how data is split into chunksets, encoded using Clay Codes into 16 chunks (10 data + 6 parity), and distributed across storage providers so any 10 of 16 chunks can reconstruct the blob. The skill covers SDK usage patterns, multipart uploads, session reuse, and token funding for APT and ShelbyUSD.

When to use it

  • Building dApps that require decentralized, durable large-object storage
  • Integrating the Shelby TypeScript SDK in Node.js or browser apps
  • Uploading or streaming large media and training datasets (10MB+)
  • Implementing multipart uploads and efficient byte-range reads
  • Running storage provider nodes or troubleshooting RPC interactions

Best practices

  • Prefer multipart uploads for files larger than 10MB to match chunkset boundaries
  • Reuse sessions for multiple reads to reduce micropayment overhead and latency
  • Use byte-range requests for streaming and partial reads to save bandwidth
  • Design blobs immutable; use overwrite semantics rather than in-place edits
  • Monitor provider health and verify recovery by testing with minimum-chunk reconstruction

Example use cases

  • Live video streaming service that reads large segments via byte-range requests
  • Training pipeline that stores terabytes of model data and retrieves shards efficiently
  • Content delivery network using Shelby for static assets with high-read throughput
  • Archival system storing immutable datasets with long expiration timestamps
  • dApp that charges users in ShelbyUSD for bandwidth and storage with micropayment sessions

FAQ

Clay Codes provide similar storage efficiency to Reed-Solomon but require up to 4x less bandwidth during recovery, making them optimal for large-scale distributed storage.

What are blob naming constraints?

Blob names are <account>/<user-defined-path>, up to 1024 characters, must not end with '/', and the namespace is flat—use paths only as identifiers.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational