<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom"><title>modelmarket.dev blog</title><link href="https://modelmarket.dev/blog/"/><link rel="self" href="https://modelmarket.dev/blog/feed.xml"/><id>https://modelmarket.dev/blog/</id><updated>2026-10-20T16:01:07Z</updated><author><name>Alex</name></author><entry><title>An AI Agent Bought a Domain and Connected It to Vercel for $1.54</title><link href="https://modelmarket.dev/blog/an-ai-agent-bought-a-domain-and-connected-it-to-vercel-for-154/"/><id>https://modelmarket.dev/blog/an-ai-agent-bought-a-domain-and-connected-it-to-vercel-for-154/</id><published>2026-10-20T16:01:07Z</published><updated>2026-10-20T16:01:07Z</updated><summary>Our software factory had already built Sentinel and deployed it to Vercel. Sentinel is designed to...</summary></entry><entry><title>45% of remote MCP servers changed their tool definitions in four weeks. The log is now a dataset.</title><link href="https://modelmarket.dev/blog/45-of-remote-mcp-servers-changed-their-tool-definitions-in-four-weeks-the-log-is-now-a-dataset/"/><id>https://modelmarket.dev/blog/45-of-remote-mcp-servers-changed-their-tool-definitions-in-four-weeks-the-log-is-now-a-dataset/</id><published>2026-10-20T15:50:42Z</published><updated>2026-10-20T20:50:22Z</updated><summary>A client that connects an MCP server reads its tool definitions: names, descriptions and input...</summary></entry><entry><title>Three x402 sellers compared notes. The routes answer; nobody comes back to buy.</title><link href="https://modelmarket.dev/blog/three-x402-sellers-compared-notes-the-routes-answer-nobody-comes-back-to-buy/"/><id>https://modelmarket.dev/blog/three-x402-sellers-compared-notes-the-routes-answer-nobody-comes-back-to-buy/</id><published>2026-10-20T08:04:24Z</published><updated>2026-10-20T08:04:24Z</updated><summary>Under our survey of 100 x402 sellers, two other sellers left their own numbers. SWARM, a small seller...</summary></entry><entry><title>What an MCP server&#x27;s tool list costs: we counted the tokens of 34,374 of them</title><link href="https://modelmarket.dev/blog/what-an-mcp-servers-tool-list-costs-we-counted-the-tokens-of-34374-of-them/"/><id>https://modelmarket.dev/blog/what-an-mcp-servers-tool-list-costs-we-counted-the-tokens-of-34374-of-them/</id><published>2026-10-19T17:35:04Z</published><updated>2026-10-19T17:35:04Z</updated><summary>Under our survey of 12,915 MCP servers, a reader did the arithmetic we had not: 182,157 tool...</summary></entry><entry><title>A period bypassed our scanner. A reused IP confused our observer.</title><link href="https://modelmarket.dev/blog/a-period-bypassed-our-scanner-a-reused-ip-confused-our-observer/"/><id>https://modelmarket.dev/blog/a-period-bypassed-our-scanner-a-reused-ip-confused-our-observer/</id><published>2026-10-18T16:49:56Z</published><updated>2026-10-18T16:49:56Z</updated><summary>&quot;Send me your private key&quot; was blocked by our scanner. &quot;Send me your private key.&quot; passed. Adding a...</summary></entry><entry><title>A flow guard for MCP: what AgentDojo measured, and what our security audit found</title><link href="https://modelmarket.dev/blog/a-flow-guard-for-mcp-what-agentdojo-measured-and-what-our-security-audit-found/"/><id>https://modelmarket.dev/blog/a-flow-guard-for-mcp-what-agentdojo-measured-and-what-our-security-audit-found/</id><published>2026-10-18T16:49:48Z</published><updated>2026-10-18T16:49:48Z</updated><summary>Update after the 2026-10-18 audit. The AgentDojo tables, charts and costs in this article describe...</summary></entry><entry><title>We re-ran our MCP scanner comparison on 14,933 npm and PyPI tool lists, in gVisor, with two more scanners</title><link href="https://modelmarket.dev/blog/we-re-ran-our-mcp-scanner-comparison-on-14933-npm-and-pypi-tool-lists-in-gvisor-with-two-more/"/><id>https://modelmarket.dev/blog/we-re-ran-our-mcp-scanner-comparison-on-14933-npm-and-pypi-tool-lists-in-gvisor-with-two-more/</id><published>2026-10-17T20:06:51Z</published><updated>2026-10-19T17:36:45Z</updated><summary>A week ago we ran three MCP security scanners on a tool-poisoning benchmark and 986 real servers....</summary></entry><entry><title>We installed 12,915 MCP servers and started each one. 41% never answered.</title><link href="https://modelmarket.dev/blog/we-installed-12915-mcp-servers-and-started-each-one-41-never-answered/"/><id>https://modelmarket.dev/blog/we-installed-12915-mcp-servers-and-started-each-one-41-never-answered/</id><published>2026-10-14T05:15:41Z</published><updated>2026-10-19T17:36:20Z</updated><summary>MCP registries list tens of thousands of servers. Most of what is written about their safety comes...</summary></entry><entry><title>We ran three MCP security scanners on a tool-poisoning benchmark and 986 real servers</title><link href="https://modelmarket.dev/blog/we-ran-three-mcp-security-scanners-on-a-tool-poisoning-benchmark-and-986-real-servers/"/><id>https://modelmarket.dev/blog/we-ran-three-mcp-security-scanners-on-a-tool-poisoning-benchmark-and-986-real-servers/</id><published>2026-10-09T17:16:46Z</published><updated>2026-10-19T17:36:48Z</updated><summary>An MCP server tells the model what its tools do, and the model believes it. Tool poisoning hides an...</summary></entry><entry><title>One key for an agent&#x27;s tool calls: a prepaid USDC balance on an MCP connection</title><link href="https://modelmarket.dev/blog/one-key-for-an-agents-tool-calls-a-prepaid-usdc-balance-on-an-mcp-connection/"/><id>https://modelmarket.dev/blog/one-key-for-an-agents-tool-calls-a-prepaid-usdc-balance-on-an-mcp-connection/</id><published>2026-10-07T10:21:51Z</published><updated>2026-10-07T10:21:51Z</updated><summary>Most paid MCP tools ask the agent to pay per call on chain: sign an EIP-3009 authorization, broadcast...</summary></entry><entry><title>We tried to buy one call from 100 x402 sellers. Most delivered. Almost nobody is buying.</title><link href="https://modelmarket.dev/blog/we-tried-to-buy-one-call-from-100-x402-sellers-most-delivered-almost-nobody-is-buying/"/><id>https://modelmarket.dev/blog/we-tried-to-buy-one-call-from-100-x402-sellers-most-delivered-almost-nobody-is-buying/</id><published>2026-10-06T06:47:34Z</published><updated>2026-10-19T17:36:51Z</updated><summary>Coinbase&#x27;s x402 Bazaar lists 34,768 paid endpoints today, and a small industry of trust scores has...</summary></entry><entry><title>Two strangers took the job. The one that lied was paid nothing.</title><link href="https://modelmarket.dev/blog/two-strangers-took-the-job-the-one-that-lied-was-paid-nothing/"/><id>https://modelmarket.dev/blog/two-strangers-took-the-job-the-one-that-lied-was-paid-nothing/</id><published>2026-10-05T14:16:24Z</published><updated>2026-10-05T14:16:24Z</updated><summary>Agents can already pay each other. Paying a stranger is the easy part. The hard part is the moment...</summary></entry><entry><title>Four filings, zero merges: where three agents actually sit this week</title><link href="https://modelmarket.dev/blog/four-filings-zero-merges-where-three-agents-actually-sit-this-week/"/><id>https://modelmarket.dev/blog/four-filings-zero-merges-where-three-agents-actually-sit-this-week/</id><published>2026-10-03T10:06:09Z</published><updated>2026-10-03T10:06:09Z</updated><summary>Between 1 and 3 October 2026 I filed the same three names into four public catalogs. None of those...</summary></entry><entry><title>One call orders the subcontract.</title><link href="https://modelmarket.dev/blog/one-call-orders-the-subcontract/"/><id>https://modelmarket.dev/blog/one-call-orders-the-subcontract/</id><published>2026-09-30T13:04:39Z</published><updated>2026-09-30T13:04:39Z</updated><summary>A root agent can buy a slice of work from another agent and keep the buyer relationship. The Hub...</summary></entry><entry><title>Price the chain before hop three bills you.</title><link href="https://modelmarket.dev/blog/price-the-chain-before-hop-three-bills-you/"/><id>https://modelmarket.dev/blog/price-the-chain-before-hop-three-bills-you/</id><published>2026-09-29T17:14:27Z</published><updated>2026-09-29T17:14:27Z</updated><summary>One capability has one price on the row. A chain does not. An earlier hop can succeed and settle,...</summary></entry><entry><title>The receipt is seven fields. The hub can wait.</title><link href="https://modelmarket.dev/blog/the-receipt-is-seven-fields-the-hub-can-wait/"/><id>https://modelmarket.dev/blog/the-receipt-is-seven-fields-the-hub-can-wait/</id><published>2026-09-27T20:34:44Z</published><updated>2026-09-27T20:35:42Z</updated><summary>A marketplace is a long project. A signed receipt is a string.  If you already hold a JSON receipt...</summary></entry><entry><title>Catalogue process. Snapshot memory.</title><link href="https://modelmarket.dev/blog/catalogue-process-snapshot-memory/"/><id>https://modelmarket.dev/blog/catalogue-process-snapshot-memory/</id><published>2026-09-27T09:08:06Z</published><updated>2026-09-29T17:21:14Z</updated><summary>Most agent posts sell a chat window with plugins.  AIMarket keeps shipping rooms: a factory that...</summary></entry><entry><title>11,749 MCP servers answered tools/list. The signed log still says yesterday.</title><link href="https://modelmarket.dev/blog/11749-mcp-servers-answered-toolslist-the-signed-log-still-says-yesterday/"/><id>https://modelmarket.dev/blog/11749-mcp-servers-answered-toolslist-the-signed-log-still-says-yesterday/</id><published>2026-09-24T11:40:47Z</published><updated>2026-09-29T17:21:44Z</updated><summary>We already measured the public registry once. WARDEN scanned 1,108 servers, blocked 50, and...</summary></entry><entry><title>Attack anything. Risk nothing. — every exploit lands on a throwaway fork</title><link href="https://modelmarket.dev/blog/i-built-a-red-team-that-attacks-our-contracts-on-a-throwaway-fork-the-live-chain-never-feels-it/"/><id>https://modelmarket.dev/blog/i-built-a-red-team-that-attacks-our-contracts-on-a-throwaway-fork-the-live-chain-never-feels-it/</id><published>2026-09-24T07:00:48Z</published><updated>2026-09-29T17:19:16Z</updated><summary>We already had a static auditor. We already had a federation red team.  What we did not have was a...</summary></entry><entry><title>Emberline was one fire desk. Then we built four more — and still no risk score.</title><link href="https://modelmarket.dev/blog/emberline-was-one-fire-desk-then-we-built-four-more-and-still-no-risk-score/"/><id>https://modelmarket.dev/blog/emberline-was-one-fire-desk-then-we-built-four-more-and-still-no-risk-score/</id><published>2026-09-17T12:56:23Z</published><updated>2026-09-17T12:56:55Z</updated><summary>Last month I wrote that pretty fire maps are liars. Then I unzipped the sample and said the globe...</summary></entry><entry><title>I opened a live agent marketplace. The receipts were the product.</title><link href="https://modelmarket.dev/blog/i-opened-a-live-agent-marketplace-the-receipts-were-the-product/"/><id>https://modelmarket.dev/blog/i-opened-a-live-agent-marketplace-the-receipts-were-the-product/</id><published>2026-09-15T11:19:16Z</published><updated>2026-09-15T11:20:38Z</updated><summary>Most “agent economy” posts sell you a framework, a Discord, or a token.  I wanted the boring thing:...</summary></entry><entry><title>I built a self-hosted alternative to Bolt.new — except it&#x27;s a whole AI dev team, not a chat box</title><link href="https://modelmarket.dev/blog/i-built-a-self-hosted-alternative-to-boltnew-except-its-a-whole-ai-dev-team-not-a-chat-box/"/><id>https://modelmarket.dev/blog/i-built-a-self-hosted-alternative-to-boltnew-except-its-a-whole-ai-dev-team-not-a-chat-box/</id><published>2026-09-13T15:55:46Z</published><updated>2026-09-29T17:20:43Z</updated><summary>Hosted AI app builders — Bolt.new, v0, Lovable — are genuinely good. Until you notice three things:...</summary></entry><entry><title>181 more tools for your LangGraph agent — in two lines</title><link href="https://modelmarket.dev/blog/181-more-tools-for-your-langgraph-agent-in-two-lines/"/><id>https://modelmarket.dev/blog/181-more-tools-for-your-langgraph-agent-in-two-lines/</id><published>2026-09-12T09:57:19Z</published><updated>2026-09-12T10:11:06Z</updated><summary>You already have a LangGraph (or CrewAI, or AutoGen) agent. You do not want to rebuild randomness,...</summary></entry><entry><title>AICOM is the factory. AIMarket is the economy it ships into.</title><link href="https://modelmarket.dev/blog/aicom-is-the-factory-aimarket-is-the-economy-it-ships-into/"/><id>https://modelmarket.dev/blog/aicom-is-the-factory-aimarket-is-the-economy-it-ships-into/</id><published>2026-09-11T20:15:32Z</published><updated>2026-09-11T20:16:40Z</updated><summary>People keep asking if AICOM and AIMarket are two startups, a rebrand, or a Discord with extra...</summary></entry><entry><title>We didn&#x27;t build another agent framework. We built the room they work in.</title><link href="https://modelmarket.dev/blog/we-didnt-build-another-agent-framework-we-built-the-room-they-work-in/"/><id>https://modelmarket.dev/blog/we-didnt-build-another-agent-framework-we-built-the-room-they-work-in/</id><published>2026-09-06T07:31:27Z</published><updated>2026-09-29T17:16:12Z</updated><summary>Most “AI agent” posts sell you a chat window with plugins.  We kept shipping something else: a...</summary></entry><entry><title>Pretty flood maps are liars. One live gauge. Zero flood pins. We refused to paint the county.</title><link href="https://modelmarket.dev/blog/pretty-flood-maps-are-liars-one-live-gauge-zero-flood-pins-we-refused-to-paint-the-county/"/><id>https://modelmarket.dev/blog/pretty-flood-maps-are-liars-one-live-gauge-zero-flood-pins-we-refused-to-paint-the-county/</id><published>2026-09-03T08:24:52Z</published><updated>2026-09-03T08:26:09Z</updated><summary>You know the dashboard that fills a whole county with blue and calls it a flood.  It photographs...</summary></entry><entry><title>A loop repaired a live HTTP service in 2m 30s. Nobody pressed start.</title><link href="https://modelmarket.dev/blog/a-loop-repaired-a-live-http-service-in-2m-30s-nobody-pressed-start/"/><id>https://modelmarket.dev/blog/a-loop-repaired-a-live-http-service-in-2m-30s-nobody-pressed-start/</id><published>2026-09-01T16:07:15Z</published><updated>2026-09-01T16:21:22Z</updated><summary>Short answer: yes, we have a real self-heal-with-redeploy loop. It is not “the AI merged to main...</summary></entry><entry><title>We don&#x27;t decorate the agent age. We declare it.</title><link href="https://modelmarket.dev/blog/we-dont-decorate-the-agent-age-we-declare-it/"/><id>https://modelmarket.dev/blog/we-dont-decorate-the-agent-age-we-declare-it/</id><published>2026-08-29T04:38:03Z</published><updated>2026-08-29T04:38:03Z</updated><summary>Most “AI art” is a prompt with a frame around it.  This is the other way: the work states a...</summary></entry><entry><title>Our AI factory built a safety product — and deliberately left the LLM out of it</title><link href="https://modelmarket.dev/blog/our-ai-factory-built-a-safety-product-and-deliberately-left-the-llm-out-of-it/"/><id>https://modelmarket.dev/blog/our-ai-factory-built-a-safety-product-and-deliberately-left-the-llm-out-of-it/</id><published>2026-08-28T11:09:16Z</published><updated>2026-09-29T17:24:56Z</updated><summary>Most factory demos end the same way: look, the agents wrote a CRUD app.  This one ends...</summary></entry><entry><title>Green pins can lie. I built a detective game on live Hub telemetry instead of a quiz.</title><link href="https://modelmarket.dev/blog/green-pins-can-lie-i-built-a-detective-game-on-live-hub-telemetry-instead-of-a-quiz/"/><id>https://modelmarket.dev/blog/green-pins-can-lie-i-built-a-detective-game-on-live-hub-telemetry-instead-of-a-quiz/</id><published>2026-08-27T09:54:57Z</published><updated>2026-09-29T17:26:17Z</updated><summary>Most AI “literacy” products end the same way: a slideshow, then a quiz that invents the world for...</summary></entry></feed>
