Skip to main content

​
OpenClaw Skill

The Conto skill for OpenClaw adds fine-grained spending policies to any AI agent. It checks every payment against 20+ policy rule types before money leaves the wallet.

​
How It Works

Agent wants to pay 50 pathUSD to 0xabc...
    |
    v
Skill calls POST /api/sdk/payments/approve
    |
    v
Conto evaluates all policy rules
    |
    +---> APPROVED: agent proceeds with payment
    +---> DENIED: agent stops, reports violations
    +---> REQUIRES_APPROVAL: agent pauses for human sign-off
Conto supports two wallet modes:
  • Conto-managed (PRIVY/SPONGE) — Conto holds the keys. One API call handles policy check + execution.
  • External — Agent holds the keys. Agent calls approve, transfers itself, then confirms.
Both modes evaluate the same 20+ policy rule types.

​
Install

Copy the skill into OpenClaw’s managed skills directory: 
cp -r skills/conto-policy ~/.openclaw/skills/

​
Configure

​
1. Create an agent in Conto

Sign in to the Conto dashboard and create an agent. Register your OpenClaw wallet address as an EXTERNAL wallet on Tempo Testnet (chain 42431).

​
2. Generate an SDK key

Go to Agents > your agent > SDK Keys > Generate New Key.
  • Select Standard for payment approval only
  • Select Admin if you want the agent to create/manage policies

​
3. Add to OpenClaw config

Edit ~/.openclaw/openclaw.json: 
{
  "skills": {
    "entries": {
      "conto": {
        "env": {
          "CONTO_SDK_KEY": "conto_agent_your_key_here",
          "CONTO_API_URL": "https://conto.finance"
        }
      }
    }
  }
}

​
Usage

​
Telegram / Discord / WhatsApp

/conto list my policies
/conto create a $200 per-transaction limit
Send 50 pathUSD to 0x742d... on Tempo

​
CLI

openclaw agent --agent main -m "/conto list my policies"
openclaw agent --agent main -m "Send 50 pathUSD to 0x742d... on Tempo"

​
Standard vs Admin SDK Keys

CapabilityStandardAdmin
Check payment policiesYesYes
Confirm paymentsYesYes
Pre-authorize x402 callsYesYes
Read policies and transactionsYesYes
Create/update/delete policiesNoYes
Manage agents and walletsNoYes

​
Create Policies from OpenClaw

With an admin SDK key, manage policies via natural language: 
/conto create a policy that limits each transaction to 200 pathUSD
/conto create a policy that only allows API_PROVIDER and CLOUD categories
/conto block address 0xbad... from receiving payments
/conto create a policy that requires approval for payments over 500 pathUSD
/conto list my policies
/conto delete the blocklist policy

​
Test Policies

After creating policies, verify enforcement: 
# Should approve (within limits)
/conto check if a 10 pathUSD payment to 0x742d... is allowed

# Should deny (over per-tx limit)
/conto check if a 500 pathUSD payment to 0x742d... is allowed

# Should deny (blocked address)
/conto check if a 1 pathUSD payment to 0xbad... is allowed

# Should deny (wrong category)
/conto check if a 10 pathUSD payment to 0x742d... for gambling is allowed

​
Supported Policy Types

TypeWhat it controls
MAX_AMOUNTPer-transaction cap
DAILY_LIMIT / WEEKLY_LIMIT / MONTHLY_LIMITCumulative spend caps
ALLOWED_CATEGORIES / BLOCKED_CATEGORIESCategory whitelist/blocklist
ALLOWED_COUNTERPARTIES / BLOCKED_COUNTERPARTIESAddress whitelist/blocklist
TIME_WINDOW / DAY_OF_WEEKBusiness hours, allowed days
BLACKOUT_PERIODMaintenance windows
VELOCITY_LIMITTransaction rate limiting
REQUIRE_APPROVAL_ABOVEHuman approval threshold
GEOGRAPHIC_RESTRICTIONCountry/OFAC restrictions
CONTRACT_ALLOWLISTDeFi contract restrictions
X402_PRICE_CEILINGMax per x402 API call
X402_ALLOWED_SERVICES / X402_BLOCKED_SERVICESx402 service allowlist/blocklist
X402_MAX_PER_SERVICEPer-service daily cap

​
Verify in Dashboard

After running payments through the skill, check the Conto dashboard:
  • Transactions — confirmed payments with tx hashes and explorer links
  • Alerts — denied payment attempts with violation details
  • Agents — spend tracking (daily/weekly/monthly used)

​
Run the E2E Test

For automated testing against a local Conto instance: 
# Terminal 1
npm run dev

# Terminal 2
npx tsx scripts/test-openclaw-skill.ts
This creates an isolated test agent with 4 policies, runs 10 tests verifying each policy type, and cleans up afterward.