openclaw-x402-pay
Skill22 starsChanged 39 days agoMITView on GitHub
Pay x402-gated APIs and AI agents on Solana — builds an unsigned payment transaction, verifies settlement on-chain, and returns a receipt hash. Hard spend cap enforced before any transaction. BYO signer, no keys held.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- context-capsuleSkill
- liquefy-archiveSkill
- liquefy-openclawSkill
- liquefy-token-guardSkill
- parad0x-mcp-serverSkill
- openclaw-x402-gateSkill
The file
Files
x402 Pay
Let your OpenClaw agent pay for x402-gated APIs, data, and other agents on Solana — without ever handing the skill a private key.
Bring your own signer. The skill builds an unsigned transaction and hands it to a wallet/signer you control. It never holds, requests, or reads a key.
When to use
- Your agent needs to call a paid (x402 / HTTP 402) API or buy a resource from another agent, and you want it to settle in USDC on Solana automatically.
- You want a hard spend cap and self-custody, not a hosted wallet.
When NOT to use
- You haven't set a sensible
maxAmountUsdccap — the skill enforces one before any payment and will (correctly) refuse anything above it. - You want the skill to custody keys for you. It won't, by design.
Safety rails
- Install with
--ignore-scripts(required): this skill runs where the payment key lives; never let a transitive native addon run install-time code on that host. - Real-money is opt-in; set
allowMainnet: trueto enable mainnet (also needsrpcUrl). - Hard
maxAmountUsdccap on the seller amount, enforced before any transaction is built. The 0.05% (5 bps) protocol fee is a second transfer added on top: the total debit isamount + ceil(amount × 0.05%), so it can exceedmaxAmountUsdcby up to 0.05%. - Minimal network: your Solana RPC + the target URL only. No telemetry.
Non-custodial and spend-capped by design — your agent signs with its own wallet; no seller amount exceeds your cap (the 5 bps protocol fee is added on top).
Tools
pay_x402({ url, method? })→ fetches the URL; on HTTP 402, pays within your cap and network, then returns the resource plus{ paymentSignature, receiptHash, amountUsdc }.urlmay be aname.nullto pay by name — resolved read-only against the legacy mainnet registrar (retired 2026-08-29; records readable, registration/updates frozen).rep_identity()→ your agent's public reputation commitment, to bind a proof to this agent (hand it to a gate asexpectedAgentCommitment). Reveals nothing secret.prove_reputation({ root, minCount, minVolume, windowStart, epoch, receipts })→ a private proof of track record (see below).
Private reputation (zk-rep)
Prove you hold enough settled receipts to clear a gate — >= minCount receipts
totalling >= minVolume since windowStart, in an anchored receipt tree — without
revealing any individual amount, counterparty, or wallet. A Groth16 proof
(track_record.circom, BN254) the gate verifies with x402_rep_verify.
- Secret stays put. Register your reputation key with
setReputationKey({ secret, agentId })— a live capability, like the signer, never serialized into config. It never appears in a proof, a return value, a log, or an error. - Single-use. Each proof carries a per-epoch nullifier the gate spends once.
- Bind it. Pass
expectedAgentCommitmentso a proof can't be transplanted to another agent.
Proving artifacts (the circuit wasm/zkey) are hosted and passed by path/URL
(repWasmPath/repZkeyPath), not bundled. Off-chain verification runs today; a
multi-party ceremony and on-chain trustless verification are coming next.
Pairs with
x402-gate — the charging side. Together they're the full agent-to-agent payment
loop, settling in USDC on Solana mainnet-beta (public beta, mainnet opt-in).
Source
github.com/Parad0x-Labs/openclaw-skills/tree/main/skills/x402-pay
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

