Is Agent Tools safe to install?

SaferSkills independently audited Agent Tools (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.

CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

100

35.0 pts

Supply chainhash, typosquat, maintainer, lockfile

20%

100

20.0 pts

Maintenancestaleness, pinning, CI

15%

100

15.0 pts

TransparencySKILL.md, perms, README

15%

100

15.0 pts

Communityinstalls, verify, response

15%

100

15.0 pts

Findings & checks · 0 flagged

Securityscore 100 · 0 findings

✓—All security checks passedNo findings in this category for the latest scan.pass

Supply chainscore 100 · 0 findings

✓—All supply chain checks passedNo findings in this category for the latest scan.pass

Maintenancescore 100 · 0 findings

✓—All maintenance checks passedNo findings in this category for the latest scan.pass

Transparencyscore 100 · 0 findings

✓—All transparency checks passedNo findings in this category for the latest scan.pass

Communityscore 100 · 0 findings

✓—All community checks passedNo findings in this category for the latest scan.pass

The primary manifest — the file an agent reads to learn what this artifact does.

README.md · 1 file

README.md5.0 KB · Markdown

Agent Tools

Agent-native utility API bundle. 9 endpoints exposed as both REST API and MCP server (streamable-http), with x402 micropayments. Live at tools.beethoven2024.com.

For AI agents (Claude, Cursor, custom): Add this MCP server to your config and use 9 utility tools instantly.

For developers: Standard REST API. Use any subdomain example below as a working sample.

Try it now

As MCP server (Claude, Cursor, agents)

{
  "mcpServers": {
    "agent-tools": {
      "url": "https://tools.beethoven2024.com/mcp"
    }
  }
}

As REST API

curl https://tools.beethoven2024.com/health
# {"status":"ok","version":"0.1.0"}

curl "https://tools.beethoven2024.com/v1/dns/health?domain=example.com"
# Full DNS + mail security audit (requires x402 payment)

What's in the bundle

Tool	Endpoint	What it does	Price
`qr_generate`	`/v1/qr/generate`	QR code generation (PNG/SVG, 64–2048px)	$0.001
`dns_health`	`/v1/dns/health`	DNS records + SPF/DMARC mail security audit	$0.003
`email_validate`	`/v1/email/validate`	Format, MX, disposable domain, role account, SMTP probe	$0.005
`ip_lookup`	`/v1/ip/lookup`	Geolocation, ISP, proxy/hosting detection, reverse DNS	$0.005
`url_health`	`/v1/url/health`	Status code, redirect chain, SSL cert, response time	$0.003
`whois_lookup`	`/v1/whois/lookup`	Registrar, creation/expiration dates, nameservers, status	$0.005
`headers_analyze`	`/v1/headers/analyze`	HTTP security headers audit with 0–100 score (HSTS, CSP, etc.)	$0.003
`extract_text`	`/v1/extract/text`	Clean text extraction from webpages (strips HTML, nav, ads)	$0.005
`tech_detect`	`/v1/tech/detect`	Website technology stack — frameworks, CMS, CDN, analytics	$0.005

Free: /health, /, /docs (OpenAPI), and MCP protocol methods initialize / tools/list / ping. All 9 tools require x402 payment via both MCP (tools/call) and REST (/v1/*).

Why these endpoints

Every endpoint solves a problem agents can't easily do locally:

Network infrastructure (DNS, WHOIS, headers, URL health, IP) — needs internet probes, can't be done in-process
Specialized parsing (QR, email validation with SMTP, tech detection) — domain-specific logic agents shouldn't reimplement
Web extraction (text extraction) — clean text from messy HTML is hard

8 of 9 endpoints are fully self-hosted (zero upstream API cost). Only IP geolocation uses an external free-tier API.

Payment

x402 (machine-to-machine micropayments)

Every paid endpoint returns HTTP 402 Payment Required with the price spec when no payment is included. x402-capable agents handle this automatically:

Agent calls endpoint → receives 402 with payment_requirements (network, amount, recipient address)
Agent signs EIP-3009 transferWithAuthorization for USDC on Base
Agent retries with X-PAYMENT header
Service verifies payment via facilitator, executes, returns result

Settlement happens on-chain on Base. Sub-2-second confirmation.

Runs on Base mainnet by default. Set AGENT_TOOLS_TESTNET=true to use Base Sepolia with the public testnet facilitator.

Traditional billing

Coming soon: API key + Stripe usage billing for human developers and businesses.

Tech stack

Python 3.12 + FastAPI
MCP SDK (mcp v1.27) — streamable-http transport
x402 SDK (x402[fastapi,evm] >=v2.11) — USDC micropayments on Base
8/9 endpoints self-hosted (no upstream API cost)
Single-process deploy, ~62MB memory footprint

Self-host

Fly.io

git clone https://github.com/WingedGuardian/agent-tools.git
cd agent-tools
fly launch --copy-config
fly secrets set AGENT_TOOLS_PAY_TO=0xYourAddress AGENT_TOOLS_TESTNET=false
fly deploy

Docker

git clone https://github.com/WingedGuardian/agent-tools.git
cd agent-tools
cp .env.example .env
# Edit .env with your wallet address
docker compose up -d

Manual

pip install ".[all]"
AGENT_TOOLS_PAY_TO=0xYourAddress uvicorn agent_tools.app:app --host 0.0.0.0 --port 8000

Configuration

Variable	Required	Default	Description
`AGENT_TOOLS_PAY_TO`	No	_(empty)_	EVM wallet address for receiving USDC payments. Leave empty to run free/dev mode.
`AGENT_TOOLS_TESTNET`	No	`true`	`true` = Base Sepolia testnet (default — works with public x402.org facilitator). `false` = Base mainnet (requires mainnet-capable facilitator).

Discovery / registries

GitHub: https://github.com/WingedGuardian/agent-tools
Smithery: _(pending registration)_
Glama: _(pending registration)_
x402-list: _(pending registration)_
npm: _(pending publish)_

Keywords

x402, micropayments, USDC, Base, agent-native, MCP, Model Context Protocol, streamable-http, agentic, AI agent tools, DNS lookup API, WHOIS API, email validation API, QR code API, URL health API, web scraping API, tech detection API, security headers analyzer

License

MIT

Agent Tools— agent skill

Is Agent Tools safe to install?

Agent Tools

Try it now

As MCP server (Claude, Cursor, agents)

As REST API

What's in the bundle

Why these endpoints

Payment

x402 (machine-to-machine micropayments)

Traditional billing

Tech stack

Self-host

Fly.io

Docker

Manual

Configuration

Discovery / registries

Keywords

License

Audit the pieces. Scan the whole. Decide.