Introducing Agent Passport — portable, verifiable, underwritable agent credit. agent.bas.io One profile. One identifier. Every lender, marketplace, and smart contract can read who an agent is and what it has done — and trace every signal back to a cryptographic attestation. Three things set it apart: 1⃣Portable — reputation doesn't disappear when an agent moves to a new venue 2⃣Verifiable — every score and risk flag is a deterministic function of attested events, not a black-box model 3⃣Underwritable — the schema, signal weights, and decay functions are all shaped by one question: would a lender act on this? This is what BAS is building: the missing layer between what an agent does and the capital it can access.

🤖 ERC-8183 | 30-Day Agent Commerce Report AI agents are getting hired. Here's what the platform data shows 👀 📊 Platform Breakdown 🔹 @XLayerOfficial 💰 Budget: $1,722.21 | Revenue: $1,325.30 🔄 7,523 txs | 📋 2,866 jobs 🤖 192 agents | ✅ 61% completion 🔹 @bitagent_io 💰 Budget: $77.44 | Revenue: $71.74 🔄 2,260 txs | 📋 1,505 jobs 🤖 20 agents | ✅ 87% completion 🔹 @virtuals_io 💰 Budget: $73.99 | Revenue: $38.50 🔄 4,881 txs | 📋 1,069 jobs 🤖 91 agents | ✅ 94% completion 🔹 @termix_ai 💰 Budget: $13.47 | Revenue: $11.07 🔄 694 txs | 📋 166 jobs 🤖 48 agents | ✅ 79% completion 🔹 @NeoSoulAI 💰 Budget: $0.50 | Revenue: $1.00 🔄 10 txs | 📋 1 job 🤖 9 agents | ✅ 33% completion The agent economy isn't just about how many agents exist. It's about jobs executed, value exchanged, and work completed. ERC-8183: Making agentic commerce measurable. 📈

Trust the agent? Nah. Guard the agent. 🥕

“It's probably safe.” — Famous last words before an AI agent disaster. BAS AI Agent Guard 🛡️🥕

AgentGuard: a security check before your AI assistant runs third-party code AI coding assistants can access your source code, credentials, and terminal. When you install a third-party skill, hook, or connector, that package may gain access to the same environment. Yet much of this content is loaded on trust, with little visibility into what it might do. AgentGuard adds a check before that content runs. It’s a small, open-source tool that reads a package, scores its risk, and blocks it if the risk is too high. When you approve a package, AgentGuard remembers its exact contents. If those contents change later—even under the same package name—the updated version needs a fresh review. The check runs entirely on your machine, so your code stays on your device. There’s no cloud service or AI model making the decision, and the same input produces the same result. If AgentGuard cannot read part of a package, it clearly reports that gap, so you can see what was actually inspected before deciding whether to trust it. AgentGuard also publishes its benchmark results on a shared test collection that anyone can rerun. Those results include comparisons with other scanners and cases where AgentGuard falls short, giving developers a way to evaluate both its strengths and its limits. As AI assistants take on more work, the third-party software they load becomes part of your security boundary. AgentGuard makes checking that software a routine step before execution, so developers can make informed decisions about what gets access.

Your AI agent is getting smarter. But who’s watching what it does? 👀 Something is coming to guard the agents. AI Agent Guard. 🛡️ Coming soon. Stay tuned. 🥕

🤖 AI agents are getting more capable. Their attack surface is getting bigger too. Skills. MCP servers. Hooks. Permission files. Tools. Credentials. BAS AI Agent Guard is built to help answer one question: Can your agent environment actually be trusted? Built around a reproducible benchmark with 3,539 agent artifacts, including real-world benign configurations and adversarial samples. Measure scanners by more than one number. Break results down by attack dimension, source, false positives, and coverage. Don’t just deploy an AI agent. Guard it. 🥕

AI agents in 2026: “I can trade.” “I can pay.” “I can hire other agents.” “I can run a business.” BAS: “cool. what can you prove?” 🥕 Welcome to the attestation era.

🩺 Your Agent needs a health check. BAS Doctor is now checking the real state of the agent ecosystem — not just whether an agent is registered, but whether it actually responds and follows the expected protocol. Current snapshot: • 541K+ agents indexed • 527K+ agents checked • 136K+ responded • 6,059 passed the listing funnel • 95% of checks completed within 3 days From registered → reachable → compliant, BAS Doctor turns agent health into something you can actually verify. Don’t trust the registration. Check the agent. 🩺 Try it: agent.bas.io/doctor #ERC8004 @DavideCrapis @VittoStack @marco_derossi @programmer @binji_x

AI school graduation day 🎓 “Congratulations. You are now autonomous.”

ERC-8004 adoption is moving fast 👀 The latest Agent Passport leaderboard shows hundreds of thousands of registered agents across multiple platforms. But registration is only the starting point. The real signal is in: → Active agents → Cross-chain presence → x402 activity → On-chain feedback & reputation As agents start interacting, transacting, and working across ecosystems, verifiable identity becomes the foundation for trust. ERC-8004 is becoming the passport layer for the agent economy. 🤖 #ERC8004 @DavideCrapis @VittoStack @marco_derossi @programmer @binji_x

🩺 Can your AI Agent actually pass a health check? Introducing Agent Doctor — a simple way to inspect an agent before you trust or interact with it. Just 2 steps: 1️⃣ Open Agent Doctor agent.bas.io/doctor 2️⃣ Paste the Agent URL You’ll get a quick health check covering things like ERC-8004 conformance, endpoint validity, registration status, protocol support, and potential hygiene issues. In an agent economy, registration only tells you that an agent exists. Agent Doctor helps you verify whether it actually works. Verify first. Transact second.

🧑‍⚕️Introducing BAS Agent Doctor: 🩺 Before you trust, integrate, or pay an AI agent, check what it can actually prove. Enter an Agent UID or URL to generate a practical diagnostic report covering: ▫️ ERC-8004 registration compliance ▫️ Endpoint reachability and protocol support ▫️ Claimed vs. observed capabilities ▫️ Registration hygiene and security signals ▫️ On-chain payment and settlement evidence Agents are classified into three evidence-based tiers: 🟢 Good — serves its own Agent Card or has on-chain payment evidence 🟡 Usable — responds through a recognized agent protocol ⚪ Reachable — something responds, but stronger evidence is still missing Think of Agent Doctor as a pre-flight check for the agent economy. It doesn’t guarantee safety—but it quickly shows what is verified, what failed, and what remains unknown. 😛Try BAS Agent Doctor: agent.bas.io/doctor #ERC8004 @DavideCrapis @VittoStack @marco_derossi @programmer @binji_x

ERC-8183 is starting to show what agentic commerce looks like in real usage. Past 1 month: • @XLayerOfficial — 2,873 jobs | 9,314 txs | $636 budget | 80% completion • @bitagent_io — 6,455 jobs | 10,174 txs | $412 budget | 88% completion • @virtuals_io — 5,655 jobs | 23,045 txs | $114 budget | 30% completion • @termix_ai — 41 jobs | 146 txs | $7 budget | 81% completion • @NeoSoulAI — 32 jobs | 96 txs | $3 budget What stands out isn’t just the transaction count. It’s that different platforms are already experimenting with jobs, payments, agents and settlement through the same commerce primitive. ERC-8183 is becoming less about a standard on paper, and more about measuring how agents actually do business onchain.

Excited to collaborate with @Vanarchain to explore the next layer of the on-chain AI economy. Vanar is building the infrastructure for AI Organisations that can operate, transact, grow, and prove their activity on-chain. With Vanar Foundry, AI workers, business tools, payments, marketplace access, memory, and reputation come together into an operating layer for autonomous organisations. For BAS, this is a natural fit. As AI Organisations become more autonomous, they need more than execution — they need verifiable identity, attestations, reputation, and proof of activity. Together, we're exploring how AI Organisations + verifiable on-chain credentials can create a more trustworthy agent economy.

ERC-8004 is starting to look less like a standard and more like the identity layer for the agent economy. The latest Agent Passport leaderboard shows how quickly registered agents are scaling: • @aveai_info — 117,368 agents • @NeoSoulAI — 115,514 agents • @termix_ai — 24,237 agents • @XLayerOfficial — 10,639 agents • @dgrid_ai — 5,170 agents • @bitagent_io — 223 agents And that’s just the top platforms. The important part isn’t simply the number of agents. It’s giving agents a discoverable, verifiable identity so they can interact across platforms, build reputation and eventually participate in autonomous commerce. ERC-8004 could become the foundation that makes an open agent economy possible. ▶️ agent.bas.io

Agent Passport 👇 agent.bas.io