🛡️ Non-Custodial Security & Policy Firewall

The Execution Firewall for
Autonomous Solana Agents

Autonomous AI agents signing raw Solana transactions risk catastrophic capital loss from hallucinations, toxic slippage spikes, and DEX reverts (0x1771). Gatekeeper sits between your agent's decision and the wallet signature — guaranteeing deterministic execution, capital protection, and zero gas burned on failed trades.

🔑 Get Free API Key ⚡ Top-Up Credits OpenAPI Specs

Deterministic Security & Policy Architecture

Gatekeeper creates an impenetrable barrier between AI agent logic and on-chain capital allocation.

🛡️

Anti-Hallucination Guardrails

Enforces strict max trade sizes, slippage bounds, and token balance integrity. If the agent hallucinates a toxic swap, Gatekeeper drops it instantly.

⚡

Zero-Gas Revert Intercept

Off-chain pre-flight simulations against real-time cluster state detect DEX reverts (0x1771) and pool locks in <80ms before any transaction is broadcast.

🔒

100% Non-Custodial Security

Private keys NEVER touch Gatekeeper. Your agent keeps its keys in its own secure environment — Gatekeeper evaluates unsigned intents objectively.

🎯

Dynamic CU Clamping

Automatically clamps over-provisioned 200,000 CU swap budgets down to the simulated requirement (* 1.12x), saving ~150k CU per trade.

# 1. Install via PyPI
pip install gatekeeper-solana

# 2. Protect autonomous agents in 5 lines of code
import asyncio
from gatekeeper_solana import GatekeeperAsyncClient

async def main():
    gk = GatekeeperAsyncClient(api_key="gk_production_key", base_url="https://gk.ai-futures-bot.pro")
    verdict = await gk.evaluate_intent(
        token_in="So11111111111111111111111111111111111111112",    # SOL
        token_out="EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v",   # USDC
        amount_in=100_000_000,                                    # 0.1 SOL
        max_slippage_bps=50,
    )
    if verdict.is_approved:
        print(f"✅ Approved in {verdict.execution_time_ms}ms! Clamped CU: {verdict.clamped_compute_units}")
    else:
        print(f"🛡️ HARD_ABORT: {verdict.rejection_reason} - 0 Gas Burned!")

asyncio.run(main())

// 1. Install via npm
npm install gatekeeper-solana-sdk

// 2. Evaluate before broadcast
import { GatekeeperClient } from "gatekeeper-solana-sdk";

const gk = new GatekeeperClient({
  apiKey: process.env.GATEKEEPER_API_KEY!,
  baseUrl: "https://gk.ai-futures-bot.pro"
});

const verdict = await gk.evaluateIntent({
  token_in: "So11111111111111111111111111111111111111112",
  token_out: "EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v",
  amount_in: 100000000,
  max_slippage_bps: 50,
});

if (verdict.decision === "APPROVED") {
  console.log(`✅ Approved: Clamped CU ${verdict.clamped_compute_units}`);
} else {
  console.warn(`🛡️ Hard Abort: ${verdict.rejection_reason} - Saved: ${verdict.fees_saved_lamports} lamports`);
}

# Direct HTTP Evaluation via High-Speed Gateway
curl -s -X POST https://gk.ai-futures-bot.pro/api/v1/preflight \
  -H "Content-Type: application/json" \
  -H "X-Gatekeeper-Key: gk_production_key" \
  -d '{
    "token_in": "So11111111111111111111111111111111111111112",
    "token_out": "EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v",
    "amount_in": 100000000,
    "max_slippage_bps": 50
  }'

💬

Developer & Community Feedback

We are actively developing Gatekeeper in the open to protect autonomous agents across Solana. We'd love to hear from developers running trading bots and agent systems:

• What failure modes are you seeing most often? (slippage spikes, blockhash expiration, account contention)
• Would a pre-flight execution firewall fit into your current trading pipeline?
• What guarantees or benchmarks would you need to see before trusting it with a live bot?

🐛 Share Feedback on GitHub 🤝 Superteam Profile

The Execution Firewall forAutonomous Solana Agents

Deterministic Security & Policy Architecture

Free Tier: 10,000 Pre-Flight Checks / Month

Developer & Community Feedback

The Execution Firewall for
Autonomous Solana Agents