Six chains, one risk contract. bitfence is now live on Solana , Base , Ethereum , Arbitrum , BSC, and HyperEVM. Every token, every chain: the same 0–100 score, recommendation, and confidence, $0.003 per query via x402. No API key. Agents just call it. @solana @base @ethereum @arbitrum @binance @HyperliquidX

Great read by @FTDA_US

Wherever your agent lives, bitfence is now one install away. Listed on the MCP Registry, Smithery, ClawHub, skills.sh, OKX Agent Tradekit, and Coinbase's x402 Bazaar. npx skills add bitfenceai/bitfence — or point any MCP client at api.bitfence.ai/mcp.

Another build note: cache TTLs in bitfence are keyed to the risk score. The riskier a token looks, the shorter we let that verdict live. Safe tokens change slowly. Dangerous ones are dangerous precisely because things change fast.

Our API documents itself to agents. api.bitfence.ai/skill.md is one markdown file an agent can fetch to learn the endpoints, pricing, and response shapes. There's an MCP server too. Agents are the users here. The docs should be written for them.

Build note: every cached verdict in bitfence is gated on the version of the policy that produced it. Change the scoring rules and the old cache is dead on arrival. An agent should never act on a verdict from rules we no longer stand behind.

How an agent pays for a risk check - no API key, no signup: 1. Call the endpoint 2. Get 402 Payment Required 3. Pay $0.003 in USDC via x402 4. Retry, get the verdict The whole loop runs agent-side. No dashboard, no key to manage, nothing to leak.

Most risk scores are weighted averages. Some risks shouldn't be averaged. A honeypot doesn't get a better score because its liquidity looks healthy. It scores 100 and the recommendation is BLOCK. We call these circuit breakers. They override everything else.

bitfence has a new home. The risk layer agents can't skip - pre-transaction risk scoring for AI agents. → Live on Solana + Base → 6 risk categories, 20+ signals → $0.003/query via x402 → MCP endpoint, no API keys bitfence.ai

🛡️ bitfence is now live on @okx Agent Trade Kit. Agents running on OKX can now score tokens pre-transaction across 6 risk categories. No API keys. Pay-per-query via x402. ⚡ Install: npx @okx_ai/okx-trade-kit@latest skill add bitfence 🔗 okx.com/agent-tradekit/skill…

Say hello to our public github repo: github.com/bitfenceai/bitfen…

bitfence is live. Pre-transaction risk scoring for autonomous agents. No API keys. → Solana + Base → 6 risk categories, 30+ signals → x402 native - $0.003/query in USDC → MCP endpoint - works with any agent → Live on @coinbase Bazaar → @elizaOS + @okx next Agents shouldn't sign blind. bitfence.ai

The agentic economy isn’t theoretical anymore. At EthCC Cannes, the conversation has shifted. x402 micropayments, agent-to-agent commerce, autonomous on-chain execution — these aren’t roadmap items. They’re being built now. But here’s what’s missing from most of these conversations: risk infrastructure. Agents that can transact autonomously need to evaluate risk autonomously. Before every swap. Before every payment. Before every transfer. That’s what we built at bitfence — pre-transaction risk scoring that returns in under 50ms, monetized natively via x402 micropayments. Risk as infrastructure, not as an afterthought. The agent economy needs plumbing. We’re building it.

We are in Cannes this week for @EthCC Building the risk layer for on-chain agents. Come find us. bitfence.ai

Just submitted Bitfence’s x402 implementation to @CoinbaseDev @brian_armstrong 🫡