Reported issues for ClawPay MCP — Non-Custodial x402 Payments for AI Agents

Pod holds 6 of 6 problems reported by people outside the maintainer team. Issues filed by the project's own owners, members and collaborators are excluded entirely — a maintainer's release checklist is not a warning to a prospective user.

Back to ClawPay MCP — Non-Custodial x402 Payments for AI Agents.

Most discussed

Question: signed receipts for selected AgentPay approval / spend / payment handoffs?

Hi — I found agentpay-mcp while looking at MCP-native payment/spend-governance projects where agents can hit paid APIs, request approval, enforce spend limits, simulate payments, and produce records that may later be used for review, accounting, or dispute workflows.

I’m building BoundaryAttest, a small open-source project for portable signed receipts around selected claims, actions, artifacts, or handoffs that cross trust boundaries:

https://github.com/cullenmeyers/BoundaryAttest

The idea is

Read the thread · 2026-08-25 · open · external user · 0 comments

Bounded buyer/seller interoperability test with a live x402 v2 endpoint

Hi — I operate Maha Strategies' live x402 v2 seller endpoint for Context Compression and would like to test AgentPay MCP as an independently maintained buyer.

Proposed bounded test:

POST https://www.mahastrategies.com/api/v1/compress
Base Mainnet (eip155:8453), USDC
Exact ceiling: 1,000 USDC base units ($0.001)
Exercise AgentPay MCP's challenge parsing and spend-policy decision
Issue at most one signature and one paid request
Make no automatic paid retry
Verify the settlement

Read the thread · 2026-08-18 · open · external user · 0 comments

Example value-packs: withRetry(fetchWithPayment) can settle twice on a timed-out response, and the budget policy misses the first payment

Hi — I work on exactly-once execution for agent payments, and I read the value-pack examples alongside the SDK. I think the retry composition in the shipped examples can pay twice for one logical fetch, and it's worth fixing there because the examples are exactly what integrators copy.

The chain (all in examples/value-packs/):

shared/x402-client.ts — fetchWithPayment runs with AbortSignal.timeout(30000). An x402 settlement can complete after that abort fires: the authorization

Read the thread · 2026-08-17 · open · external user · 0 comments

Daily Review FAILED — 2026-08-15

Daily Review FAILED for 2026-08-15 (UTC): the workflow errored before producing its report issue.

Failed run: https://github.com/up2itnow0822/agentpay-mcp/actions/runs/31902330937

Likely causes (see RUNBOOK.md troubleshooting):

CLAUDE_CODE_OAUTH_TOKEN expired or the subscription is inactive. Tokens last about a year and do not auto-renew: re-run 'claude setup-token' locally and reset the secret.
The github_token: line was removed from the claude-code-action step. Scheduled runs fail the OI

Read the thread · 2026-08-15 · open · external user · 24 comments

clawpay-mcp is indexed on our MCP marketplace

Hey -- found clawpay-mcp while indexing servers for https://getlulu.dev/mcps. Non-custodial x402 payments for AI agents. Sign locally, spend limits, Base network.

You're indexed and ranked well already:

Servers in this category are seeing real usage-based monetization interest through our SDK -- worth a look if you'd like to turn usage into revenue. Otherwise just a heads-up, badg

Read the thread · 2026-08-10 · open · external user · 0 comments

Design question: non-custodial flow vs MCP transport, payment-attempt observability, x402 receipt persistence

Hi — I'm working on OM World, a protocol for a decentralized intent economy. agentpay-mcp's framing — non-custodial x402 MCP payment layer as the open alternative to closed offerings — overlaps directly with our Agent Mandate (authority to spend) and Execution Proof (receipt of spend) primiti

Read the thread · 2026-05-17 · open · external user · 1 comment

Most recent

The remaining reports are on the project's issue tracker.