Solana Clawd Agents
Use this skill when working with the imported Solana Clawd Agents catalog at agents/.
It covers agent JSON definitions, catalog generation, API publishing, MCP discovery,
Metaplex Core minting, staking support, x402 paid-agent wiring, Cloudflare API
deployment, and localized public catalog assets.
Source Snapshot
- Local source imported from
/Users/8bit/agents/agents
- Hub copy lives at
solana-clawd-agents/agents
- Primary catalog:
agents/agents-catalog.json
- Manifest:
agents/agents-manifest.json
- Agent definitions:
agents/src/*.json
- Public API snapshot:
agents/public/api/agents/
- Build entrypoint:
agents/build-catalog.cjs
Read Order
- Read
agents/AGENTS.md for repo-local operating guidance.
- Read
agents/README.md for the catalog, API, minting, staking, and x402 map.
- For API or deploy work, read
agents/docs/API.md and agents/docs/DEPLOYMENT.md.
- For x402 work, read
agents/docs/X402_IMPLEMENTATION.md.
- For a specific agent, read
agents/src/<agent-id>.json and compare it with
agents/agents-catalog.json.
Workflows
- Catalog edits: modify
agents/src/*.json, then run the local catalog build in
agents/.
- Catalog verification: run
npm test from agents/ when Node is available.
- Public API work: compare generated files under
agents/public/api/agents/
with agents/agents-catalog.json.
- Templates: use
agents/agent-template.json,
agents/agent-template-full.json, or agents/agent-template-attested.json
as the starting point.
- Minted examples: inspect
agents/minted/*.json before changing mint-related
metadata.
- Character definitions: inspect
agents/characters/*.json for personality or
runtime profile changes.
Safety Boundaries
- Do not launch, mint, stake, bind tokens, or register identities on mainnet
without explicit confirmation.
- Do not store private keys, seed phrases, raw wallet signatures, or wallet
approval payloads in agent definitions.
- Keep agent permissions bounded: explicit tools, spend caps, wallet authority,
model/provider, memory policy, and runtime status.
- Use devnet, dry-run, or validation-only commands first for minting, staking,
and paid-agent flow changes.
1---2name: solana-clawd-agents-23description: Use the imported Solana Clawd Agents hub catalog to browse, validate, build, deploy, mint, stake, and integrate production-ready Solana agent definitions, including x402 paid-agent APIs, MCP catalog endpoints, Cloudflare Worker API, Metaplex Core minting, and agent JSON schema workflows.4---56# Solana Clawd Agents78Use this skill when working with the imported Solana Clawd Agents catalog at `agents/`.9It covers agent JSON definitions, catalog generation, API publishing, MCP discovery,10Metaplex Core minting, staking support, x402 paid-agent wiring, Cloudflare API11deployment, and localized public catalog assets.1213## Source Snapshot1415- Local source imported from `/Users/8bit/agents/agents`16- Hub copy lives at `solana-clawd-agents/agents`17- Primary catalog: `agents/agents-catalog.json`18- Manifest: `agents/agents-manifest.json`19- Agent definitions: `agents/src/*.json`20- Public API snapshot: `agents/public/api/agents/`21- Build entrypoint: `agents/build-catalog.cjs`2223## Read Order24251. Read `agents/AGENTS.md` for repo-local operating guidance.262. Read `agents/README.md` for the catalog, API, minting, staking, and x402 map.273. For API or deploy work, read `agents/docs/API.md` and `agents/docs/DEPLOYMENT.md`.284. For x402 work, read `agents/docs/X402_IMPLEMENTATION.md`.295. For a specific agent, read `agents/src/<agent-id>.json` and compare it with30 `agents/agents-catalog.json`.3132## Workflows3334- Catalog edits: modify `agents/src/*.json`, then run the local catalog build in35 `agents/`.36- Catalog verification: run `npm test` from `agents/` when Node is available.37- Public API work: compare generated files under `agents/public/api/agents/`38 with `agents/agents-catalog.json`.39- Templates: use `agents/agent-template.json`,40 `agents/agent-template-full.json`, or `agents/agent-template-attested.json`41 as the starting point.42- Minted examples: inspect `agents/minted/*.json` before changing mint-related43 metadata.44- Character definitions: inspect `agents/characters/*.json` for personality or45 runtime profile changes.4647## Safety Boundaries4849- Do not launch, mint, stake, bind tokens, or register identities on mainnet50 without explicit confirmation.51- Do not store private keys, seed phrases, raw wallet signatures, or wallet52 approval payloads in agent definitions.53- Keep agent permissions bounded: explicit tools, spend caps, wallet authority,54 model/provider, memory policy, and runtime status.55- Use devnet, dry-run, or validation-only commands first for minting, staking,56 and paid-agent flow changes.