Base MCP agent safety studio

Base AgentGuard

Plan AI wallet actions before approval: inspect Base RPC health, score agent intent risk, and export a Base MCP send_calls-ready receipt.

Network

Chain ID
--
Latest block
--
Gas price
--
Status
Waiting

Intent builder

Describe the agent action

Risk score

0 Not scored

Findings

    Base MCP receipt

    send_calls preview

    {}

    Base fit

    Built around Base MCP

    Mirrors the docs pattern: detect wallet, prepare unsigned calldata, map it into send_calls, and require user approval for every write.

    OSS fit

    Maintainer-first workflows

    Includes a plugin spec, contract sketch, examples, roadmap, and contribution docs so it can become a real public repo.

    Airdrop posture

    Useful Base activity

    Focuses on building, documenting, testing, and deploying on Base instead of relying on fake claim links or speculative shortcuts.

    Evidence pack

    What reviewers should see immediately

    Base MCP Wallet detection, send_calls mapping, approval-first UX
    x402 HTTPS endpoint review, max USDC cap, external-data warnings
    Builder Codes Agent attribution plan for Base.dev analytics and discovery
    Base deploy Sepolia proof contract, Basescan verification checklist
    OpenAI OSS Maintainer workflow, issue triage, PR review, release automation
    Trust No custody, no seed phrase, no hidden execution, bounded approvals