Agent Skills for BlockVectra: multi-chain EVM RPC, the Data API, programmatic sign-up and pricing.
Keyless RPC availability, methods and limits follow each chain's public policy in GET /v1/chains. The Data API requires an API key.
One command (uses the skills CLI (npm package skills); add -g for a user-level install, -a claude-code to pick the agent, -y to skip prompts):
npx skills add blockvectra/skillsOr as a Claude Code plugin, together with the docs MCP server:
/plugin marketplace add blockvectra/mcp
/plugin install blockvectra-docs@blockvectra
| Skill | Use it for |
|---|---|
blockvectra-rpc |
EVM JSON-RPC and Data API calls, per-chain public access and limits, wallet-signature sign-up, error handling, pricing lookup and USDT/USDC top-up |
blockvectra-signup |
Obtain a BlockVectra API key without a browser: wallet-signature (SIWE/EIP-191) sign-up, first-login account creation with free CU grant, create and revoke API keys programmatically |
The skill reads chains, limits and prices from live endpoints instead of copying numbers, so it does not go stale.
Skills are directories with a SKILL.md file. Per the Agent Skills overview:
-
Claude Code: place the skill directory in
~/.claude/skills/(personal) or.claude/skills/(project). See Use Skills in Claude Code.git clone https://github.com/blockvectra/skills.git mkdir -p ~/.claude/skills cp -r skills/blockvectra-rpc ~/.claude/skills/
-
claude.ai: upload the skill as a zip file under Settings > Features (paid plans with code execution enabled; see the overview above for details). Zip the
blockvectra-rpcfolder. -
Claude API: upload through the Skills API (
/v1/skillsendpoints). Skills on the API run without network access, so the live lookups in this skill need an environment with network access, such as Claude Code.
Other agents that read SKILL.md directories can use the same folder; check that tool's documentation for its skills path.
claude mcp add --transport http blockvectra-docs https://docs.blockvectra.com/mcpMIT