Overview
Goldsky Agent is a skill pack for AI coding assistants that helps you build, deploy, and debug across the full Goldsky product surface — Turbo pipelines, Mirror pipelines, Subgraphs, Compose, and Edge RPC.
Skills auto-trigger based on what you describe — no slash commands required. Interactive workflow skills walk you through multi-step tasks end-to-end. Reference skills provide YAML syntax, manifest fields, error patterns, and CLI commands on demand.
Works with Claude Code, Cursor, Windsurf, OpenCode, Codex, and other AI coding assistants.
Quick start
| I want to… | Use |
|---|
| Build a new Turbo pipeline | /turbo-builder |
| Fix a broken Turbo pipeline | /turbo-doctor |
| Fix a broken Mirror pipeline | /mirror-doctor |
| Fix a broken Compose app | /compose-doctor |
| Build / deploy a subgraph | /subgraph-builder |
| Fix a broken / stalled subgraph | /subgraph-doctor |
| Migrate a subgraph from The Graph | /subgraph-migrate |
| Build a Compose app (oracle / keeper / automation) | /compose |
| Get a fast, reliable RPC endpoint | /edge |
| Find the right dataset name | /datasets |
| Look up Turbo YAML syntax | /turbo-pipelines |
| Look up Compose manifest, CLI flags, or TaskContext | /compose-reference |
| Set up the CLI and log in | /auth-setup |
Installation
npx (recommended)
Claude Marketplace
Claude Code (local)
Cursor (local)
Manual
Installs to Claude Code, Cursor, and 30+ other AI agents:npx skills add goldsky-io/goldsky-agent
/plugin marketplace add goldsky-io/goldsky-agent
/plugin install goldsky@goldsky-agent
git clone https://github.com/goldsky-io/goldsky-agent.git
claude --plugin-dir ./goldsky-agent
Clone the repo and add the path to Cursor settings (Settings > Cursor Settings > JSON):git clone https://github.com/goldsky-io/goldsky-agent.git
{
"plugins.local": ["/absolute/path/to/goldsky-agent"]
}
git clone https://github.com/goldsky-io/goldsky-agent.git
cp -r goldsky-agent/skills/* .claude/skills/ # Claude Code
cp -r goldsky-agent/skills/* .cursor/skills/ # Cursor
Use with autonomous or BYO agents
Editors like Claude Code and Cursor auto-discover skills and trigger them from each skill’s description. Autonomous or bring-your-own-model agents — Devin, Codex, custom LLM apps, or a plain ChatGPT/Claude chat — don’t implement that convention. Even after you install the skill files, these agents won’t find them on their own: npx skills add copies the skills into a skills directory (.claude/skills/ for a single Claude Code target, or the shared .agents/skills/ when you target several agents) but does not add any entry-point file telling an agent to look there. Use Option 1, Option 2, or both.
Option 1 — Connect the docs MCP (recommended)
The hosted Goldsky docs MCP lets any MCP-capable agent search all Goldsky docs and discover the skills (exposed as MCP resources) at query time — no auto-trigger convention needed.
Point your client at https://docs.goldsky.com/mcp. Clients that use the common mcpServers schema (e.g. Cursor, Claude Code) take:
{
"mcpServers": {
"goldsky": { "url": "https://docs.goldsky.com/mcp" }
}
}
servers key; Claude.ai adds it as a custom connector) — see Goldsky MCP server for per-client setup. Devin is MCP-compatible via a mcpServers entry in .devin/config.json; see Devin’s MCP docs for the exact schema (a hosted HTTP server may need a small mcp-remote bridge).
Option 2 — Install the skill files, then point your agent at them
# From the docs site (also serves skills at /.well-known/skills/):
npx skills add https://docs.goldsky.com
# Or from the skills repo:
npx skills add goldsky-io/goldsky-agent
.claude/skills/<name>/ or .agents/skills/<name>/ — each with SKILL.md + references/. Because no entry-point file is created, add an AGENTS.md at your project root so agents that read it (Devin, Codex, Amp, …) know the skills exist (point the paths at wherever your install landed):
# AGENTS.md
This project uses Goldsky. AI-agent skills for building, deploying, and
debugging Goldsky pipelines are installed under `.agents/skills/`.
Before working on a Goldsky task (Turbo / Mirror pipelines, Subgraphs,
Compose, Edge RPC), read the matching `.agents/skills/<name>/SKILL.md`
and the files it references under `.agents/skills/<name>/references/`.
For reference lookups (CLI flags, YAML fields, dataset names), query the
Goldsky docs MCP: https://docs.goldsky.com/mcp
These agents don’t trigger skills automatically — tell the agent to consult the relevant skill or the docs MCP for the task (e.g. “use the Goldsky Turbo transforms skill to decode this Solana instruction”).
Available skills
Skills are grouped by product. Each group has interactive workflow skills (guided, multi-step) and/or reference skills (lookup-oriented).
Turbo pipelines
Streaming pipelines that index onchain data from 130+ chains into PostgreSQL, ClickHouse, Kafka, S3, and more.
| Skill | When to use | What it does |
|---|
turbo-builder | ”I want to build a pipeline for X” | Guides you through chain → dataset → transforms → sink → validate → deploy |
turbo-doctor | ”My pipeline is broken / not getting data / output looks wrong” | Diagnoses the problem step-by-step and offers to run fixes |
turbo-pipelines | ”What’s the YAML syntax for X? Should I use dataset or Kafka?” | Source/transform/sink field reference and architecture decisions (flow patterns, sizing, sink selection) |
turbo-transforms | ”How do I decode EVM logs / write a SQL transform?” | SQL, TypeScript/WASM, dynamic tables, HTTP handlers |
turbo-operations | ”How do I pause / restart / delete? What does this error mean?” | Lifecycle commands, pipeline states, CLI monitoring, error patterns |
Mirror pipelines
Goldsky’s original streaming pipeline product. Prefer Turbo for new pipelines — reach for Mirror only when you need a subgraph entity source, the one thing Turbo can’t do.
| Skill | When to use | What it does |
|---|
mirror | ”How do I sync my subgraph to PostgreSQL? Mirror vs Turbo?” | Sources, sinks, lifecycle commands, Mirror vs Turbo guidance |
mirror-doctor | ”My Mirror pipeline is failing / stuck / terminated” | Runs status and log commands, identifies root cause, applies fixes |
Subgraphs
Hosted GraphQL APIs for dApp frontends. For most other use cases, prefer Turbo — it’s faster, cheaper, and more reliable. Use subgraphs when you specifically need a GraphQL endpoint.
| Skill | When to use | What’s inside |
|---|
subgraph-builder | ”Build / write / deploy a subgraph; design a schema; write a mapping” | Interactive author→build→deploy; schema design, AssemblyScript mappings, manifest, instant subgraphs, performance, testing; endpoints, tags, webhooks |
subgraph-doctor | ”My subgraph stopped syncing / won’t deploy / is throwing errors” | Diagnostic workflow: status + log checks, the _meta query, error-pattern matching, preventive mapping-code root causes, fixes |
subgraph-migrate | ”Move my subgraph off The Graph onto Goldsky” | Path selection, deploy, verify sync, migrate tags, swap endpoint |
Compose
Offchain-to-onchain TypeScript framework for oracles, keepers, circuit breakers, and cross-chain automation.
| Skill | When to use | What it does |
|---|
compose | ”Build a price oracle / keeper / cross-chain bot in TypeScript” | Walks through scaffolding, task triggers (cron, HTTP, onchain), wallets, gas sponsorship |
compose-doctor | ”My Compose app is in error state / crashlooping / not processing tasks” | Runs status, logs, secret list, wallet list and diagnoses |
compose-reference | ”What fields does compose.yaml accept? What’s the TaskContext API?” | Manifest fields, every goldsky compose flag, TaskContext / wallet / Collection APIs, codegen, pricing |
Edge (managed RPC)
Globally distributed, low-latency JSON-RPC for EVM chains, built on eRPC with intelligent routing, caching, and failover.
| Skill | When to use | What’s inside |
|---|
edge | ”RPC rate limits, hedged requests, flashblocks, x402, error code -32005” | Capabilities, supported chains, pricing, dashboard, error code reference |
Cross-cutting
Used across multiple products.
| Skill | When to use | What it does |
|---|
auth-setup | ”Install the CLI / log in / switch projects / fix unauthorized errors” | Walks through CLI installation, login, and project switching |
secrets | ”Create credentials for PostgreSQL / ClickHouse / Kafka / webhook sinks” | Guides credential creation and secret management |
datasets | ”What’s the dataset name for Polygon NFTs? What prefix does Solana use?” | Chain prefixes, dataset types, naming conventions |
How it works
Skills are automatically discovered by your AI assistant and triggered based on what you describe. You don’t need to invoke them by name.
User: "Build me a pipeline for USDC transfers on Base"
↓
turbo-builder (auto-triggered)
↓ references
turbo-pipelines + datasets + secrets
↓
Generated pipeline.yaml + deployment
turbo-builder, turbo-doctor, mirror-doctor, compose-doctor) run as agents with access to CLI tools. Reference skills provide structured knowledge that the AI uses to answer questions or generate YAML.
Repository structure
goldsky-agent/
├── skills/
│ ├── turbo-builder/ # Step-by-step pipeline creation wizard
│ ├── turbo-doctor/ # Diagnose and fix Turbo pipeline issues
│ ├── turbo-pipelines/ # YAML configuration + architecture reference
│ ├── turbo-transforms/ # SQL, TypeScript, dynamic tables
│ ├── turbo-operations/ # Lifecycle, monitoring, error patterns
│ ├── mirror/ # Mirror pipeline deploy, operate, reference
│ ├── mirror-doctor/ # Diagnose and fix Mirror pipelines
│ ├── subgraph-builder/ # Author, build & deploy subgraphs; schema/mappings/manifest
│ ├── subgraph-doctor/ # Diagnose and fix failing/stalled subgraphs
│ ├── subgraph-migrate/ # Guided migration from The Graph
│ ├── compose/ # Compose app scaffolding, triggers, wallets
│ ├── compose-doctor/ # Diagnose and fix Compose apps
│ ├── compose-reference/ # compose.yaml fields, CLI flags, TaskContext API
│ ├── edge/ # Managed RPC capabilities, error codes, pricing
│ ├── datasets/ # Chain prefixes, dataset types
│ ├── secrets/ # Credential management
│ └── auth-setup/ # CLI installation, login
├── hooks/
│ └── scripts/ # Validation, secret checking
└── .claude-plugin/ # Plugin manifest
Pre-deploy hooks
When installed as a plugin, Goldsky Agent runs hooks automatically on goldsky turbo apply commands:
| Hook | What it does |
|---|
pre-deploy-validate | Runs goldsky turbo validate, blocks on failure |
secret-check | Verifies all secret_name references exist |
post-deploy-inspect | Suggests goldsky turbo inspect after deploy |
Coverage
The skills cover the full Goldsky product surface:
- Turbo pipelines — 130+ chain sources (EVM, Solana, Bitcoin, Stellar, Sui, NEAR, Starknet); SQL / TypeScript / dynamic table transforms; PostgreSQL, ClickHouse, Kafka, S3, Webhook, S2, SQS, MySQL, Pub/Sub sinks; streaming and job modes; full lifecycle and monitoring
- Mirror pipelines — Subgraph and direct-indexing sources, sinks, lifecycle, plus interactive diagnosis
- Subgraphs — Author/build/deploy (
subgraph-builder: schema design, AssemblyScript mappings, manifest, instant subgraphs, performance, testing, endpoints/tags/webhooks); interactive diagnosis (subgraph-doctor); guided migration from The Graph (subgraph-migrate)
- Compose —
compose.yaml manifest, cron / HTTP / onchain triggers, smart wallets, gas sponsorship, TaskContext API, codegen, pricing
- Edge RPC — Capabilities, supported chains, hedged requests, flashblocks, x402, error code lookups
- Cross-cutting — Authentication, secrets, dataset naming, full CLI reference
Example prompts
Once installed, describe what you need in natural language:
- “Build a Turbo pipeline to track USDC transfers on Base”
- “My pipeline is stuck in error state — help me fix it”
- “Sync my subgraph entities to PostgreSQL with a Mirror pipeline”
- “Build a subgraph for an ERC-721 collection on Ethereum”
- “Migrate my subgraph from The Graph to Goldsky”
- “My subgraph stopped syncing — help me figure out why”
- “Build a Compose task that runs a price oracle every 5 minutes”
- “My Compose app is returning 500 on the HTTP trigger — what’s wrong?”
- “Why am I getting RPC error -32005 on Edge?”
- “What’s the dataset name for Polygon ERC-20 transfers?”
- “Show me the YAML syntax for a ClickHouse sink”
- “Configure a PostgreSQL secret for my pipeline”
- “Set up the Goldsky CLI and authenticate”
