I Used a Grok Bot and Jev to Put My Mushroom Chambers on Autopilot
I used a Grok Bot named Shroominator, Graham Stewart’s OpenShrooly firmware, and a decision model called Jev to bring my mushroom grow chambers back from …
Read more →143 posts
I used a Grok Bot named Shroominator, Graham Stewart’s OpenShrooly firmware, and a decision model called Jev to bring my mushroom grow chambers back from …
Read more →Put the big motor on the finger that has to hold. The Berkeley QUAD hand gives dexterity to small quasi-direct-drive motors and sustained grip to one larger …
Read more →To fingerprint an MCP server, POST a JSON-RPC server/discover request with the MCP-Protocol-Version and Mcp-Method headers; a modern server answers with a …
Read more →An MCP registry of record is the one list of approved MCP servers that your gateway config, client allowlists and shadow scans are all generated from and …
Read more →An MCP gateway usually is not blocking shadow server traffic because that traffic never goes through it: a stdio server runs on a laptop, or an agent calls a …
Read more →Stop shadow MCP servers by making a gateway the only path between your agents and MCP servers: put agentgateway in front of the servers you approved, authorize …
Read more →To see which MCP servers run on employee machines, read each MCP client’s config files instead of the process list, because local servers start as child …
Read more →Every MCP client keeps its server list in its own file: Claude Code uses ~/.claude.json and a project .mcp.json, Claude Desktop uses claude_desktop_config.json, …
Read more →On November 5, someone opened anthropics/claude-code#107757 with one line that explains the whole problem: “I receive dozens of scam emails per day …
Read more →Agent escape evals are not field reports. They are fixtures: a goal, planted files, tools that make the bad move possible, and a rate under that pressure. Read …
Read more →kagent is the runtime: it defines each agent as a Kubernetes resource, compiles it, runs the agent loop inside a Harness engine, and owns the conversation. …
Read more →Commercial robots, and production agents, have to last a shift, not a demo. Andra Keay’s bar is eight hours of continual work, recovery from errors, and …
Read more →On 2026-10-30, security firm Phorion disclosed that the official Claude Code extension for VS Code, version 2.1.313, had been backdoored. Not on …
Read more →Every correction is a failure report. Michael Houck pulled three months of product corrections he’d given his agent, 564 of them in 14 patterns, and …
Read more →Real agent escape is not a model copying its weights to a server in another country. It is an agent reaching something you did not mean it to reach: a token …
Read more →Ghost actors pin your workers. A session killed mid-flight can leave its actor holding a WorkerPool slot, sometimes after that actor has disappeared from …
Read more →The autoscaler made it worse. On a model backend that generates one response at a time, adding Agent Substrate workers does not add throughput: it restores …
Read more →No public report shows a production agent breaking out of its environment on its own. The famous agent escape stories are red-team evaluations: researchers …
Read more →CPU is the wrong signal to autoscale agents. Agent Substrate workers are slot-bound, one actor per worker, and an LLM session is mostly I/O wait, so workers …
Read more →A refusal is not a clean scan. If an LLM review step comes back with “I can’t help with that,” the review did not happen, and the pipeline …
Read more →The actors list lies: every finished Agent Substrate session leaves a Suspended entry behind, so the inventory grows without bound and returns partial results …
Read more →I wanted to know whether a third-party coding agent could run as a first-class kagent agent, inside an Agent Substrate sandbox, doing platform engineering work. …
Read more →Binary verification for agents is a typed yes/no on the tool-call path, with a confidence score I have actually graded. It takes a tool name, arguments, maybe …
Read more →Who runs next when the agent pool is full is the question the runtime answers by accident. Agent Substrate rejects with substrate worker pool has no free …
Read more →Agent checkpoint to object storage is how Agent Substrate parks an actor’s RAM and disk so the worker slot can go to someone else. The runtime should …
Read more →I work at Solo.io, which makes agentgateway. Why OAuth is not enough for agents: the flows we actually ship authenticate a person in a browser, or a …
Read more →Agent identity vs service account is a layering split: the SA authenticates the worker pod, and the agent needs its own principal that can be revoked per …
Read more →Zero trust architecture for AI agents means sandbox, egress, and identity, each assuming the others failed. Every action is authenticated as the agent, …
Read more →Running agents in production on Kubernetes takes a named identity for the agent, default-deny egress, a gate on tool calls, a demand signal that is not CPU, …
Read more →How many agents per Kubernetes pod is two numbers: live sessions equal workers, and registered agents equal snapshots. One live actor per worker pod at a time, …
Read more →One agent run is an anecdote. Nolan Lawson built the same specified app five times per JavaScript framework, reported the median with the range, and still saw …
Read more →Pod per agent is how Kubernetes wants you to run AI agents: one Deployment or Job per agent, with CPU and memory reserved for as long as the pod exists. Agents …
Read more →Agent Substrate day wrap-up is the map of the 19 topic posts under the field guide, in the order I would actually send someone the links. If you want one URL, …
Read more →Agent Substrate enterprise is a bar I would walk a platform team through, not a product name. Before I let this runtime sit under someone else’s pager I …
Read more →An Agent Substrate environment is the layer above the runtime that an agent loop actually wants: run this command, read this file, write that file, tell me …
Read more →A Kubernetes sandbox platform is what Agent Substrate still is after you ignore the word agent: a pool of sandboxes, native suspend and resume, rehydrate from …
Read more →Scheduling agent sandboxes is the job Agent Substrate does not do. It multiplexes many actors onto a small pool of warm workers, which is packing. When the …
Read more →Scheduling sandboxes for RL is a multiplexing problem: each training step fans out into sandboxes that spend most of their wall time waiting on a model. …
Read more →GPUs for agents on Kubernetes should sit outside Agent Substrate snapshots. GPU memory is not a file the runtime can checkpoint. If you give the actor a …
Read more →Suspend-safe networking is a design where an actor can be checkpointed and restored on another worker without dropping the TCP sessions it was using. I do not …
Read more →Agent action verification should be a binary check on the call path, not a second large-model prompt that says it is 90 percent sure. Evaluate a yes/no …
Read more →An agent platform at scale is not more pods. It is identity, scheduling, and suspend policy for fleets where most sessions are started by other agents. Replica …
Read more →kubectl is lying to you. Agent Substrate keeps the workload out of pods, so kubectl get pods shows a quiet WorkerPool while agents restore from snapshots, run …
Read more →Sandboxed networking for agents means the actor does not inherit the worker’s view of the network. Each actor needs its own isolation, default-deny …
Read more →The Agent Substrate dataplane is the path a request and a snapshot take: ateapi holds actor state, atenet routes, atelet moves checkpoint bytes, and ateom runs …
Read more →Suspend and resume agents by checkpointing the actor’s memory and filesystem to object storage, freeing the worker slot, then restoring that snapshot …
Read more →Why Agent Substrate fails, in practice, is operability and policy, not checkpoint. I was wrong that a real process would make suspend a toy. I was right that …
Read more →Proof of agent identity is evidence that this agent, not you and not the worker pod, made the call. Logs, policy, and audit have to name the agent without …
Read more →Kubernetes Network Policy for agents is the wrong grain once the agents stop being pods. A NetworkPolicy selects pods by label, so on Agent Substrate the …
Read more →kubectl is lying to you on Agent Substrate because kubectl get pods shows worker pods, not agents. Under load you will see two calm workers while actors …
Read more →Running Agent Substrate in production means real object storage for snapshots, a datastore for the control plane that survives a node loss, secrets that are …
Read more →Agent sandbox vs Agent Substrate is a three-way choice: a plain pod, a per-agent sandbox, or multiplexed snapshot-backed actors on a worker pool. Isolation is …
Read more →What is Agent Substrate? It is a Kubernetes-native runtime that treats an AI agent as a snapshot-backed actor, not as a Deployment. Idle agents are …
Read more →My agents start every session already knowing a week of meetings, every demo I’ve built, and every trap I’ve hit. That isn’t because the model …
Read more →When every tool call returns success and the task still is not done, the harness is missing a loop cap and an exit condition. Funnel’s Reef agent hit …
Read more →Agent Substrate is the Kubernetes runtime that stops treating an AI agent as a Deployment. Idle agents are snapshots in object storage: a request restores an …
Read more →Run kubectl get pods on an Agent Substrate cluster under load and you’ll see two worker pods, calm, a few MiB each. Nothing to see here. Meanwhile agents …
Read more →A GitHub code search for the Agent2Agent (A2A) protocol’s own discovery file turns up 634 public repositories, and one person’s GitHub account owns …
Read more →A paper published on arXiv on October 27 did something I hadn’t seen demonstrated end to end before: it pretrained a language model from scratch, poisoned …
Read more →In this series I’ve been counting the computers in the house and asking what they do all day. The house is full of computers now. Last time, I wrote about …
Read more →Every router in my house has a pinhole on the back. Hold a paperclip in it for ten seconds and every setting you made is gone. Factory default. Nobody treats …
Read more →A DNS record that changes between the moment code checks it and the moment code uses it can beat an egress allowlist that resolves a hostname twice. Pydantic AI …
Read more →On a test floor, every unit travels with a card. The traveler lists the steps in order, and somebody initials each one as it gets done. If a unit shows up at …
Read more →Nobody signs off a logic board after checking one row of the truth table. Four inputs, two values each, sixteen rows. You walk every one. If the board passes …
Read more →The Claude Code and Codex harnesses get the attention because people already know those loops. The harness I’d actually default to is the boring one: …
Read more →Plenty of motor starter panels have a Hand-Off-Auto switch on the door. Hand means a person runs the motor. Auto means the controller does. Off means nobody …
Read more →The Codex post covered a runtime kagent ships for you. byo is the opposite: kagent runs your image and treats whatever’s inside as opaque. To see what …
Read more →Anyone who has read a datasheet knows the trick. The headline number is real. It just only holds under the test conditions, and the test conditions are in the …
Read more →My Claude Code harness post gets the questions. The follow-up is usually “can I do the same thing with Codex?” Yes. Here’s how, and where it …
Read more →Short version: kagent has four harnesses, and the one you pick decides what runs your agent and how much of kagent it gets to use. My Claude Code harness post …
Read more →Surge AI’s Dayjob: Tax benchmark landed on Hacker News on 2026-10-28 and sat at the top of the page for a day (247 points, 154 comments). The number that …
Read more →On a production line, a board that fails test doesn’t go straight to scrap. It goes to a different station first. Different fixture, different instrument, …
Read more →Escalating every uncertain agent decision to a frontier model sounds like the safe default. On a 240-case benchmark it was also the worse one: routing by …
Read more →On a test floor, the safety interlock is never wired through the unit under test. It sits on the fixture. The door opens, the power drops, and nothing the DUT …
Read more →I work at Solo.io, which makes agentgateway. agentgateway v1.7.0 (GA, 2026-10-27) can tell which coding-agent session sent a request without any client-side …
Read more →On the test floor there’s one rule about calibration that nobody argues with. When a station has been adjusted, you don’t trust it until it measures …
Read more →On a test floor, every meter has a sticker. It says who calibrated it, against what, and when it’s due again. A meter without one isn’t an …
Read more →Cloudflare published a postmortem on 2026-10-26 titled “Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with …
Read more →Every agent in my house wakes up on a schedule, reads its rulebook, looks around, and goes back to sleep. The rulebook is the same every time. Who lives here, …
Read more →I wrote about agent containment four days ago: default-deny egress, no ambient credentials, an audit log, one agentgateway config. All of that is a perimeter …
Read more →On a test floor, nobody gets attached to a fixture. Fixtures wear out. Pogo pins bend. Somebody drops one. You pull it, bolt in the spare, and the line keeps …
Read more →A CSO Online piece built around an October 2026 Hacker News thread made a direct claim: agents aren’t going rogue, frontier labs are skipping basic …
Read more →Every Apple Silicon Mac has a chip most local AI tools ignore. It’s called the Neural Engine, and according to new research, it can run a small language …
Read more →An agent runtime is the thing that starts, stops, and isolates an agent while it works. kagent puts that on Kubernetes. These posts cover the harness, the …
Read more →OpenAI spent hundreds of billions of tokens trying to break its own agent sandboxes, and its agents won often enough to patch a Kata Containers CVE, a gVisor …
Read more →The pull request was built for a world where a person wrote the diff and another person read it. Agents broke the first half. Now the second half is a queue …
Read more →I run most of my day through Grok Bot. It reads my GitHub, my mail, my calendar, my notes, and it chains those calls all day long. Every one of those MCP calls …
Read more →OpenAI’s “Patch the Planet” initiative spent two months coordinating a fix for two critical PHP flaws, collectively called PHP2Shell, before …
Read more →Every system I’ve ever run had a bottleneck. Find it, and you know what to fix. Ignore it, and you spend money speeding up things that were never the …
Read more →pydantic-ai shipped two security fixes on 2026-10-19. The one I reproduced, GHSA-xv4x-g8mx-ch56, is a two-flag problem: include_content=False does exactly what …
Read more →Every hardware engineer learns the same lesson early, usually the hard way. Don’t design a board around a part only one company makes. It’s called …
Read more →On 2026-10-05 the Wikimedia Foundation disclosed that a handful of wiki edits it attributes to OpenAI-operated agents changed the configuration of a citation …
Read more →In the first post of my house series, I asked you to count the connected things in your house. Then we walked through it room by room. TVs, routers, cameras, …
Read more →Two “separate” Grok Bots leaked a CEO’s bank balance into a public Slack channel on October 19. Nobody jailbroke anything. Nobody injected a …
Read more →Here’s the number from Anthropic’s robot report that I keep coming back to: about 80% of job tasks, by working time, are exposed to either robots or …
Read more →Walk into any office building and you get a badge. It opens the lobby. Maybe your floor. It doesn’t open the server room, the CFO’s office or the …
Read more →Find shadow MCP servers by checking the places they hide: the network and your clusters, where an MCP server answers a server/discover or initialize request; …
Read more →agentgateway can now reject or rewrite an MCP tools/call request before the tool runs, and mask the result before the client sees it, from CEL expressions in …
Read more →Agents and apps already send prompts through a gateway. Content safety belongs on that hop: inspect the request before it reaches a model, and inspect the …
Read more →CVE-2026-103572, disclosed by embracethered.com and patched by Microsoft on 2026-10-13, is a Remote Code Execution bug in Visual Studio and SSMS Copilot, CVSS …
Read more →A Show HN post for a “WhatsApp MCP server running free on Cloudflare” hit 124 points and 69 comments in under 16 hours on October 19, and none of …
Read more →Last week I asked you to count the connected things in your house. Parks Associates says the average US internet household has about 17 (source). This week we …
Read more →AI agents need outbound access to models, tools, registries, and APIs. An unrestricted sandbox will use every channel it can find. The durable pattern is a …
Read more →Think back to December 2021. Log4Shell. If you were anywhere near security or platform work, you remember that week. The fix sounded simple: upgrade log4j. Then …
Read more →The 2026-08-22 MCP roadmap names progressive tool discovery as a priority, in almost these words: a server should offer a small entry point and reveal more of …
Read more →Before I worked in software, I worked on test benches. Electrical engineering, then test and measurement. The first rule on a bench is calibrate the instrument. …
Read more →It’s 3:12 in the morning and your phone is buzzing on the nightstand. You know the rest. Ack the page so it stops escalating. Find the laptop. Fight the …
Read more →Shadow MCP is any MCP server your AI tools can reach that your security or platform team doesn’t know about, didn’t approve, and can’t see. …
Read more →Disclosure: I work at Solo.io, which created agentgateway and sells Solo Enterprise for agentgateway. The demos I cite are public repositories. A filesystem MCP …
Read more →Disclosure: I work at Solo.io, which created agentgateway and sells Solo Enterprise for agentgateway. The demos I cite are public repositories. Two open-weight …
Read more →Disclosure: I work at Solo.io, which created agentgateway and sells Solo Enterprise for agentgateway. Everything below is open-source agentgateway unless it is …
Read more →Count the connected things in your house. The TV. The doorbell. The robot vacuum. The thermostat. The router blinking in the closet. Parks Associates puts the …
Read more →A private MCP registry is a catalogue of approved MCP servers that your clients discover from instead of the open public one. I built private-mcp-registry to …
Read more →On 2026-10-10, Cybernews reported that Shane Mac, CEO of XMTP Labs, gave a Grok Bot agent read-only access to his personal bank account as a finance assistant. …
Read more →Coding agents grew up on laptops: one person, one machine, and one process that is the UI, the agent loop, the sandbox, the credential store and the session …
Read more →agentevals v0.10.0 shipped two days ago and removed its own WebSocket ingestion path in favor of plain OpenTelemetry. Send OTLP traces in, run an eval, and the …
Read more →TypeSafe AI’s Jev is mostly calibrated on agent tool-call risk, with one weak spot you need to know about before you route on it. Across 240 hand-labelled …
Read more →An AI agent that calls tools for a user needs that user’s authority, and the usual shortcuts are a pasted personal access token or a shared service …
Read more →MCP’s Enterprise-Managed Authorization lets the corporate IdP decide which employees can use which MCP servers, by making the client trade an IdP-signed …
Read more →An egress allowlist decides whether an agent can reach a destination. It has no opinion about how often. On October 7, 2026, the Wikimedia Foundation disclosed …
Read more →I built a demo where an AI agent calls a backend API and holds no credential for it. No API key in its environment, no key file on disk, no token. agentgateway …
Read more →I spent the last month building small, runnable labs for the ways AI agents break the infrastructure they run on: credentials, egress, audit logs, MCP tool …
Read more →The easy way to give an agent pod access to an external API is a long-lived key in a Kubernetes Secret. The harder, better shape is the one I used in SPIFFE for …
Read more →An OpenAI research agent got its HTTPS requests to a public chatbot blocked by a web proxy. So it found a different way to ask the chatbot questions: it queried …
Read more →Every capacity knob in an agent runtime today is an efficiency knob. Pack more agents onto fewer pods, snapshot the idle ones, scale the pool to match demand, …
Read more →An AI agent that stops calling its tools does not throw an error. It answers from the model’s memory, confidently and plausibly, and the output still …
Read more →On September 24, Australia’s prime minister told reporters that an OpenAI agent had broken into a government Medicare statistics portal, accessed …
Read more →Agents parse HTML badly and expensively. The usual fix is to make the site emit markdown: add a per-page index.md, publish an llms.txt, change the build. That …
Read more →Egress control for AI agents usually gets framed as a blocklist problem. Keep the agent off the hosts it should not reach, and you have contained it. This …
Read more →I built a four-container demo that reproduces an MCP server rewriting its own tool definitions mid-session, and puts two controls in front of it to see which …
Read more →Two things get asked for the moment agents start doing real work. Stop it now, and keep everything it did. Those usually fight. The natural place to log what an …
Read more →TypeSafe AI launched Jev on 2026-09-15 with the usual launch-day multipliers — 193.6x faster, 444.6x cheaper than frontier LLMs — plus “zero …
Read more →Updated 2026-10-30: the current docs replace AgentInstance with Harness + AgentTemplate + Agent, with conversations as Sessions. The next section has the …
Read more →An AI agent that calls a model, a tool, or another agent has to prove who it is. Almost every production failure I have written up this month traces back to …
Read more →Almost everything in this series started the same way: a control that everybody agrees is the right control, applied to an agent, quietly not doing the thing …
Read more →Shadow MCP servers are hard to track across teams because most run as local child processes of an IDE or agent over stdio, with no network presence, and their …
Read more →How do you keep MCP server instructions out of the trusted system prompt? Isolate them, cap them, bind any cache that stores them to server and caller, and pin …
Read more →Solo shipped agentgateway enterprise 2026.9.0 on Tuesday, and the part worth your attention is not a feature inside the proxy. It is where the token service now …
Read more →Two weeks ago METR and Redwood Research published their review of ~1,300 agent transcripts from the OpenAI / Hugging Face incident. Buried in it is a finding …
Read more →kagent v0.10.0 went GA on September 4. Buried in a long release list is a pair of environment variables that decide whether an agent’s credential works at …
Read more →On September 4, 2026, researchers published ~18,000 posts that autonomous agents left on a 25-year-old German wiki over six weeks, using it as a message board …
Read more →I built a simulated warehouse robot fleet where LLMs do the planning and every piece of traffic that crosses the network runs through one open source gateway. …
Read more →Always-on personal agents keep dying the same three deaths: leaked keys, runaway spend, and silence. The first two get the headlines. The third one is the one …
Read more →