Use VeilProof from your agent

One file, skill.md, written for an agent to read. It holds the prove a wallet steps, the verify a report steps, the ten checks, the registry methods, and the demo claim as a worked example. It is the same file an AI judge reads to check a claim.

A skill is a folder with a file named SKILL.md inside it. The setup below downloads our file and saves it under that name, in the folder your agent reads.

For every project on this machine. Run this one command, then start a new session.

mkdir -p ~/.claude/skills/veilproof && curl -sL https://veilproof22.vercel.app/skill.md -o ~/.claude/skills/veilproof/SKILL.md

For one project only. Run this inside your project folder instead.

mkdir -p .claude/skills/veilproof && curl -sL https://veilproof22.vercel.app/skill.md -o .claude/skills/veilproof/SKILL.md

Check it is there with /skills in Claude Code. A skill added mid session is not picked up until you start a new one.

These agents read the same file from a different folder. In your project, save it as .agents/skills/veilproof/SKILL.md. For every project, save it as ~/.agents/skills/veilproof/SKILL.md instead.

mkdir -p .agents/skills/veilproof && curl -sL https://veilproof22.vercel.app/skill.md -o .agents/skills/veilproof/SKILL.md

Download skill.md, or give your agent this link and ask it to read the file: https://veilproof22.vercel.app/skill.md. Any agent that can read a file can follow it. That includes Muse: paste the link, or the file, and ask it to use the skill.

Your agent reads the skill when your request matches it. These all work:

Use the veilproof skill to prove my NEAR testnet wallet.

Use the veilproof skill to verify report 2e8675fa662f06a4.

Use the veilproof skill to judge this claim.

For a judge, the skill has a section with three steps to check a claim against the live registry. No wallet is needed to judge: the demo claim verifies from public data alone.