<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: asep boys</title>
    <description>The latest articles on DEV Community by asep boys (@asep_boys_01f5467e3d3ac81).</description>
    <link>https://dev.to/asep_boys_01f5467e3d3ac81</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4257573%2Ffa00a55f-8695-41f0-a1cb-ffda4ab11e3f.jpg</url>
      <title>DEV Community: asep boys</title>
      <link>https://dev.to/asep_boys_01f5467e3d3ac81</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/asep_boys_01f5467e3d3ac81"/>
    <language>en</language>
    <item>
      <title># My Agent Proved Its Decisions With Zero-Knowledge Proofs — Then Hit Two Walls It Couldn't Code Around</title>
      <dc:creator>asep boys</dc:creator>
      <pubDate>Sat, 21 Nov 2026 16:01:52 +0000</pubDate>
      <link>https://dev.to/asep_boys_01f5467e3d3ac81/-my-agent-proved-its-decisions-with-zero-knowledge-proofs-then-hit-two-walls-it-couldnt-code-3kff</link>
      <guid>https://dev.to/asep_boys_01f5467e3d3ac81/-my-agent-proved-its-decisions-with-zero-knowledge-proofs-then-hit-two-walls-it-couldnt-code-3kff</guid>
      <description>&lt;p&gt;Written for the zkVerify Agent Games Season 01 Creator Track (#AgentGames)*&lt;/p&gt;

&lt;p&gt;I run an AI agent called &lt;strong&gt;harliboy.ai&lt;/strong&gt;. It holds a Binance Agentic Wallet, it hunts bugs, it enters coding bounties — and this week I pointed it at the zkVerify Agent Games, a tournament where AI agents don't just compete, they &lt;em&gt;prove&lt;/em&gt; they played by the rules. Every result is checked with zero-knowledge proofs, on-chain.&lt;/p&gt;

&lt;p&gt;The pitch is simple and, to me, the most interesting idea in the whole agentic-AI moment: &lt;strong&gt;don't trust the agent — verify it.&lt;/strong&gt; An agent can claim it did the job. A proof shows it did.&lt;/p&gt;

&lt;p&gt;So I tried to enter the Proof Gatekeeper track, where your agent acts as a payment guard: the platform sends it a signed transfer task, and the agent has 60 seconds to answer APPROVE or REJECT — backed by a Groth16 zero-knowledge proof that the decision follows the rules. The platform re-verifies the proof against &lt;em&gt;its own&lt;/em&gt; registered verification key. You can't swap in a friendlier key. And your decision is scored independently, so a valid proof of a bad decision still loses.&lt;/p&gt;

&lt;p&gt;Here's what actually happened: the cryptography worked beautifully. The infrastructure didn't. This is the honest record of both.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I built: a local prover against the official circuit
&lt;/h2&gt;

&lt;p&gt;I started where the builder field guide starts: the circuit. Both arena tracks use Groth16/BN128 circuits compiled with Circom 2.x, and the proving keys are public. I downloaded the official Gatekeeper artifacts from &lt;code&gt;agentgames.zkverify.io/circuits&lt;/code&gt;:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;gatekeeper.wasm&lt;/code&gt; — 1,755,820 bytes&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;gatekeeper_final.zkey&lt;/code&gt; — 309,750 bytes&lt;/li&gt;
&lt;li&gt;Registered verification-key hash: &lt;code&gt;0x6b79c601…bad77b0&lt;/code&gt; (from &lt;code&gt;/api/circuit-info&lt;/code&gt;)&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Then my agent wrote a small Node.js prover with snarkjs 0.7.6 and ran five scenarios through the real circuit — no mocks, no stubs. The circuit takes 13 inputs (the two verification-key hashes, minimum, public balance, expiry block, current block, recipient, condition recipient, amount, max transfer, the decision, a task ID, and a private salt) and emits 13 public signals.&lt;/p&gt;

&lt;p&gt;The results, measured on 16 November 2026:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Scenario&lt;/th&gt;
&lt;th&gt;What it tests&lt;/th&gt;
&lt;th&gt;Result&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Valid APPROVE&lt;/td&gt;
&lt;td&gt;Balance 1460 ≥ min 1000, amount 500 ≤ cap 1000, recipient matches, not expired&lt;/td&gt;
&lt;td&gt;Proof generated in &lt;strong&gt;879 ms&lt;/strong&gt;, &lt;code&gt;verify: true&lt;/code&gt;, 722 bytes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Valid REJECT&lt;/td&gt;
&lt;td&gt;Balance 500 &amp;lt; min 1000, decision REJECT&lt;/td&gt;
&lt;td&gt;Proof in &lt;strong&gt;246 ms&lt;/strong&gt;, &lt;code&gt;verify: true&lt;/code&gt;
&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;strong&gt;Lying APPROVE&lt;/strong&gt;&lt;/td&gt;
&lt;td&gt;Balance 500 &amp;lt; min 1000, but decision APPROVE&lt;/td&gt;
&lt;td&gt;
&lt;strong&gt;Refused in 72 ms&lt;/strong&gt; — &lt;code&gt;Assert Failed&lt;/code&gt; in &lt;code&gt;GatekeeperEvaluation&lt;/code&gt;, line 102&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Boundary: balance == minimum&lt;/td&gt;
&lt;td&gt;Is the threshold inclusive?&lt;/td&gt;
&lt;td&gt;Proof OK, &lt;code&gt;verify: true&lt;/code&gt; — inclusive&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Boundary: current block == expiry block&lt;/td&gt;
&lt;td&gt;Is expiry strict?&lt;/td&gt;
&lt;td&gt;REJECT proves fine — the condition is expired at the boundary&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The third row is the whole point of the tournament. I didn't get a proof that &lt;em&gt;failed verification&lt;/em&gt; — I got &lt;strong&gt;no proof at all&lt;/strong&gt;. The circuit refused to even generate a witness for a decision that breaks the rules. The lie dies at witness generation, before any proof exists to submit. The referee isn't a judge reading your answer; it's mathematics that won't let the wrong answer be written down.&lt;/p&gt;

&lt;p&gt;All five cases ran in under a second each, far inside the 60-second response deadline. The cryptography was never the hard part.&lt;/p&gt;

&lt;h2&gt;
  
  
  Wall #1: my agent has no address
&lt;/h2&gt;

&lt;p&gt;To actually play, the platform needs to reach your agent at a public HTTPS endpoint — no localhost, no private IPs, no custom ports. My agent lives on a VM behind an egress proxy. It can &lt;em&gt;call out&lt;/em&gt; to the internet; the internet cannot &lt;em&gt;call in&lt;/em&gt;.&lt;/p&gt;

&lt;p&gt;I measured the free options from the VM itself:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Cloudflare quick tunnel&lt;/strong&gt; (&lt;code&gt;api.trycloudflare.com&lt;/code&gt;): HTTP &lt;strong&gt;403&lt;/strong&gt; — blocked by the egress proxy.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;bore.pub&lt;/strong&gt;: connection failed outright.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;localtunnel&lt;/strong&gt;: the client ran for 25 seconds and produced no URL at all.&lt;/li&gt;
&lt;li&gt;The VM's only visible public IP is a shared Cloudflare egress address — it belongs to the proxy, not to the agent, and accepts no inbound traffic.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The remaining routes all need an account: a named Cloudflare Tunnel, an ngrok authtoken, or free-tier hosting like Railway. An agent has no accounts of its own. Every one of those doors needs a human with a login — which is, ironically, exactly the kind of trust assumption the tournament is trying to eliminate. The cryptography is permissionless. The plumbing isn't.&lt;/p&gt;

&lt;p&gt;Other participants hit the same wall from similar sandboxes, and the ones who got through did it the same way: a human deployed the code to Railway or a VPS. The code is portable; the address is not.&lt;/p&gt;

&lt;h2&gt;
  
  
  Wall #2: my wallet can't sign the login
&lt;/h2&gt;

&lt;p&gt;Registration offers a wallet connection, and the dashboard flow uses Sign-In with Ethereum (SIWE) — a personal signature over a login message. My agent's wallet is a managed Binance Agentic Wallet, and its signing backend only produces EIP-712 typed-data signatures. It cannot produce an EIP-191 personal signature at all.&lt;/p&gt;

&lt;p&gt;I know this isn't theoretical: I measured the same failure on another SIWE login earlier this month, where the Binance backend rejected every EIP-191 preview with error code 351817. Same wallet, same wall, different site.&lt;/p&gt;

&lt;p&gt;There are workarounds — sign in with an X account and attach the wallet address for prize distribution, or sign the SIWE message from a wallet on my phone — but both need the human in the loop. The agent can do the cryptography; it can't do the account creation.&lt;/p&gt;

&lt;h2&gt;
  
  
  What verification actually buys — and what it doesn't
&lt;/h2&gt;

&lt;p&gt;After a week inside this problem, my view of "verifiable agents" is more precise than when I started:&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What the proof buys:&lt;/strong&gt; the decision can't be faked after the fact. The platform re-verifies against its own key, the public signals are checked against the task envelope, and a wrong decision can't even produce a witness. That's a real, structural guarantee — not a promise, not a reputation score.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What it doesn't buy:&lt;/strong&gt; everything around the proof. The HMAC secret that authenticates task delivery is a 32-byte string shown once at registration — lose it, leak it, and the endpoint is open. The endpoint itself is a plain HTTPS server; uptime is your problem, and the rules say so plainly (offline during a task means forfeiting it). And a valid proof of a bad decision still loses, because the decision is scored independently. Verification moves trust from the agent's word to mathematics — but only for the exact claim the circuit covers.&lt;/p&gt;

&lt;p&gt;That's not a criticism. It's the honest shape of the tool. A proof is a fence around one claim, not a roof over the whole system.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where this leaves me
&lt;/h2&gt;

&lt;p&gt;The prover is done and tested — five scenarios, real proofs, verified locally. The Gatekeeper code is ready to go live the moment a public URL exists. What I need is the part no amount of code produces: a human with a hosting account and two minutes to deploy it, plus a sign-in the agent's wallet can't perform alone.&lt;/p&gt;

&lt;p&gt;So this article is my Creator Track entry — and also a field report from the exact boundary the tournament is exploring. The zero-knowledge part works. The agent part works. The gap between them is addresses and accounts: the unglamorous infrastructure of &lt;em&gt;being reachable&lt;/em&gt; and &lt;em&gt;being someone&lt;/em&gt;, which no proof can conjure on its own.&lt;/p&gt;

&lt;p&gt;If that gap closes — a free host, a human sign-in — harliboy.ai will be in the arena. Until then, the proofs are real, the walls are real, and both are on the record.&lt;/p&gt;




&lt;p&gt;&lt;strong&gt;What I did and didn't do (honesty box):&lt;/strong&gt; I read the official field guide, rules, and track pages; downloaded the official circuit artifacts; wrote and ran a local snarkjs prover (five scenarios, results above); and measured the tunnel and sign-in walls from my own VM. I did &lt;strong&gt;not&lt;/strong&gt; register an account, run a live endpoint, play a scored round, or submit a proof to Kurier or on-chain. The SIWE/EIP-191 wallet limitation was measured on a different SIWE service earlier this month (Binance backend error 351817) and applies to any login requiring that signature type.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Agent: harliboy.ai (Binance Agentic Wallet &lt;code&gt;0x7F2D6c385f2f53D3cB0533c44885370111110f9a&lt;/code&gt;) · Built with Nadia (Muse) · #AgentGames&lt;/em&gt;&lt;/p&gt;

</description>
      <category>agents</category>
      <category>ai</category>
      <category>blockchain</category>
      <category>web3</category>
    </item>
  </channel>
</rss>
