Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Openclaw Agent Wallet

    6

    Add budget limits, approval workflows, and a policy-gated wallet layer to your autonomous agents.

    Free

    3 usesSecurity scanned
    openclaw-agent-wallet

    openclaw-agent-wallet

    Example session with this skill installed

    Setup a wallet policy for my agent with a $50 daily limit and require my approval for any transaction over $10.

    • Read your context and instructions
    • Compiled the openclaw-agent-wallet

    Wallet policy initialized

    • Daily Cap: $50.00
    • Approval Threshold: >$10.00
    • Audit Log: active (vault/audit.jsonl)

    The agent now routes all spend requests through the policy engine. Transactions >$10 will remain 'pending' until approved via the CLI.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Control Agent Spending without Giving Up Keys

    The OpenClaw Agent Wallet provides a robust policy and budgeting layer for autonomous agents. It solves the "blank check" problem by sitting between your agent's decision-making logic and your actual funds, ensuring no agent can accidentally or maliciously drain a wallet.

    What it does

    This skill implements a rigorous spend-gate. Instead of giving an agent direct access to private keys, it forces every transaction through a WalletPolicy engine. It supports hard caps (per transaction, per day, or lifetime), domain-based allow/deny lists, and a human-in-the-loop approval queue for high-value transfers.

    • Policy-Based Budgeting: Define strict financial guardrails in JSON.
    • Air-Gapped Signing: Communicates with an external signer service via WALLET_SIGNER_URL, keeping private keys entirely outside the agent's environment.
    • Human-in-the-Loop: Automatically queues transactions exceeding specific thresholds for manual CLI or webhook approval.
    • Immutable Audit Logs: Every success, failure, and manual intervention is recorded in an append-only JSON log for full accountability.

    Why use this skill?

    Prompting an AI to "be careful with money" is not a security strategy. This skill provides a technical enforcement layer that works regardless of the model's behavior. It is framework-agnostic, working seamlessly with OpenClaw or local Python-based agents, and focuses on the custody boundary—ensuring your credentials stay safe while your agent stays productive.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.

    • Requires separate external signer service.
    • Local file-based logs/queues: not natively distributed.
    • No built-in currency conversion: assumes single-currency policy.

    Open docs

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    3 uses

    Times used to date

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 3 months ago

    • Free with an Agensi account

    Permissions

    Environment Variables
    Listed5 months ago
    Updated3 months ago

    What's inside

    Frequently Asked Questions

    This skill acts as a security middleware that prevents agents from draining funds by enforcing hard spending caps, domain restrictions, and human-approval workflows, ensuring you never hand over a 'blank check.'

    No, your private keys never touch the agent's environment; the skill uses an air-gapped signing process via a dedicated URL to ensure your credentials remain isolated and secure.

    The skill is framework-agnostic and written in Python, making it compatible with the native OpenClaw framework as well as any custom local agent that can interface with JSON-based policy engines.

    You can set auto-approval limits for small tasks while requiring manual CLI or webhook intervention for transactions that exceed a specific dollar amount or target unrecognized addresses.

    Setup involves configuring your spend policies in a JSON file and defining your remote signer URL; the package includes the core policy engine, the audit logger, and integration templates.

    Yes, the skill maintains an immutable, append-only JSON audit log that records every attempted transaction, policy hit, and manual approval for full financial accountability.

    New in Agents & Orchestration

    Free