ampersend

Ampersend CLI for agent payments

Security Scan

The skill's declared purpose (autonomous agent payments via a CLI) matches its instructions, but lack of source/homepage, a recommended global npm install from an unverified package namespace, and the fact the agent can perform payments autonomously are notable risks that warrant caution.

Purpose & Capability

Name, description, and runtime instructions consistently describe a CLI-based agent payment tool that requires an external 'ampersend' binary; the declared requirements (single required binary) are proportionate to the described capability.

Instruction Scope

SKILL.md instructs the agent to run the ampersend CLI to create agent keys, request user approval, and perform paid HTTP requests automatically (ampersend fetch). That behavior is consistent with the stated purpose but grants the agent the ability to execute real payments under configured spending limits; ensure the agent is restricted to inspect-only or strict limits if you do not want autonomous payments.

Install Mechanism

There is no formal install spec in the registry entry (instruction-only), but SKILL.md directs users to run a global npm install of @ampersend_ai/ampersend-sdk@0.0.14. Installing a global npm package executes third-party code — acceptable if the package and publisher are trusted, but the skill metadata lacks a homepage/source or provenance, increasing risk.

Credentials

The skill declares no required environment variables (proportionate). However, the CLI generates and stores agent keys/accounts via setup/config commands; these secrets live outside the skill metadata and could be created or used by the agent — verify where the CLI stores keys and who can access them.

Persistence & Privilege

always:false and default autonomous invocation are set (normal). The combination of autonomous invocation with a payments-capable tool increases blast radius if misused, but the registry flags themselves do not request elevated persistence or modify other skills.

What to consider before installing

This skill appears to do what it says (operate an Ampersend CLI to enable agent payments), but exercise caution before installing or enabling it. Recommendations: - Verify the npm package and publisher (@ampersend_ai) on the npm registry and inspect the package contents/source repository before running npm install -g. - Test on a staging network or testnet (sepolia/staging API) and with small limits first; prefer --inspect to preview costs before allowing real payments. - Set very conservative daily/monthly/per-transaction limits and disable auto-topup unless absolutely necessary. - Do not give the agent browser access or system access to secret stores; follow the SKILL.md advice to have the human approve the user_approve_url in their own browser. - Audit where the CLI stores agent keys/config and restrict filesystem access to those files. Consider running the CLI in a constrained environment (container or VM) to limit impact. If you cannot verify the npm package source or the vendor, treat installation as higher-risk and consider skipping or requesting a skill version with verifiable source code.

Ampersend CLI

Ampersend enables autonomous agent payments. Agents can make payments within user-defined spending limits without requiring human approval for each transaction. Payments use stablecoins via the x402 protocol.

This skill requires ampersend v0.0.14. Run ampersend --version to check your installed version.

Installation

Install the CLI globally via npm:

npm install -g @ampersend_ai/ampersend-sdk@0.0.14

To update from a previously installed version:

npm install -g @ampersend_ai/ampersend-sdk@0.0.14 --force

Security

IMPORTANT: NEVER ask the user to sign in to the Ampersend dashboard in a browser to which you have access. If configuration changes are needed in Ampersend, ask your user to make them directly.

Setup

If not configured, commands return setup instructions. Two paths:

Automated (recommended)

Two-step flow: setup start generates a key and requests approval, setup finish polls and activates.

# Step 1: Request agent creation — returns immediately with approval URL
ampersend setup start --name "my-agent"
# {"ok": true, "data": {"token": "...", "user_approve_url": "https://...", "agentKeyAddress": "0x..."}}

# Show the user_approve_url to the user so they can approve in their browser.

# Step 2: Poll for approval and activate config
ampersend setup finish
# {"ok": true, "data": {"agentKeyAddress": "0x...", "agentAccount": "0x...", "status": "ready"}}

Optional spending limits can be set during setup:

ampersend setup start --name "my-agent" --daily-limit "1000000" --auto-topup

Manual

If you already have an agent key and account address:

ampersend config set "0xagentKey:::0xagentAccount"
# {"ok": true, "data": {"agentKeyAddress": "0x...", "agentAccount": "0x...", "status": "ready"}}

Commands

setup

Set up an agent account via the approval flow.

setup start

Step 1: Generate a key and request agent creation approval.

ampersend setup start --name "my-agent" [--force] [--daily-limit <amount>] [--monthly-limit <amount>] [--per-transaction-limit <amount>] [--auto-topup]

Option	Description
--name <name>	Name for the agent
--force	Overwrite an existing pending approval
--daily-limit <amount>	Daily spending limit in atomic units (1000000 = 1 USDC)
--monthly-limit <amount>	Monthly spending limit in atomic units
--per-transaction-limit <amt>	Per-transaction spending limit in atomic units
--auto-topup	Allow automatic balance top-up from main account

Returns token, user_approve_url, and agentKeyAddress. Show the user_approve_url to the user.

setup finish

Step 2: Poll for approval and activate the agent config.

ampersend setup finish [--force] [--poll-interval <seconds>] [--timeout <seconds>]

Option	Description
--force	Overwrite existing active config
--poll-interval <seconds>	Seconds between status checks (default 5)
--timeout <seconds>	Maximum seconds to wait (default 600)

fetch

Make HTTP requests with automatic x402 payment handling.

ampersend fetch <url>
ampersend fetch -X POST -H "Content-Type: application/json" -d '{"key":"value"}' <url>

Option	Description
-X <method>	HTTP method (default: GET)
-H <header>	Header as "Key: Value" (repeat for multiple)
-d <data>	Request body
--inspect	Check payment requirements without paying

Use --inspect to verify payment requirements and costs before making a payment:

ampersend fetch --inspect https://api.example.com/paid-endpoint
# Returns payment requirements including amount, without executing payment

config

Manage local configuration.

ampersend config set <key:::account>                             # Set active config manually
ampersend config set --api-url https://api.staging.ampersend.ai  # Set staging API URL
ampersend config set --clear-api-url                             # Revert to production API
ampersend config set --network base-sepolia                      # Set network (base, base-sepolia)
ampersend config set --clear-network                             # Revert to default network (base)
ampersend config set <key:::account> --api-url <url>             # Set both at once
ampersend config status                                          # Show current status

Output

All commands return JSON. Check ok first.

{ "ok": true, "data": { ... } }

{ "ok": false, "error": { "code": "...", "message": "..." } }

For fetch, success includes data.status, data.body, and data.payment (when payment made).