<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>Elroot Labs Engineering Journal</title>
<link>https://elrootlabs.com/blog</link>
<description>Field notes on AI systems, local-first products and institutional software from the team building them at Elroot Labs.</description>
<language>en</language>
<atom:link href="https://elrootlabs.com/feed.xml" rel="self" type="application/rss+xml"/>
<lastBuildDate>Fri, 20 Nov 2026 00:00:00 GMT</lastBuildDate>
<item><title>Physical AI: how robots learn in simulation first, tested in MuJoCo</title><link>https://elrootlabs.com/blog/physical-ai-simulation</link><guid isPermaLink="true">https://elrootlabs.com/blog/physical-ai-simulation</guid><pubDate>Fri, 20 Nov 2026 00:00:00 GMT</pubDate><category>robotics</category><description>Physical AI trains in simulation first. Our MuJoCo test: 7 hours of practice in 21 seconds, a controller that fell over at a 40 ms delay, and the fix.</description></item>
<item><title>Remote MCP server without a login: the guards on our public one</title><link>https://elrootlabs.com/blog/remote-mcp-server</link><guid isPermaLink="true">https://elrootlabs.com/blog/remote-mcp-server</guid><pubDate>Thu, 19 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>A remote MCP server can skip the login only if every tool is read-only and public. Here are the guards on ours, and the rate limit that failed.</description></item>
<item><title>Agentic AI, AI agents and workflows: when you need which</title><link>https://elrootlabs.com/blog/agentic-ai-vs-ai-agents</link><guid isPermaLink="true">https://elrootlabs.com/blog/agentic-ai-vs-ai-agents</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Agentic AI vs AI agents vs workflows: the sources disagree on the words. Ask who decides the next step, code or the model. Our own pipeline as the example.</description></item>
<item><title>AGENTS.md: what goes in ours, and the bug behind each rule</title><link>https://elrootlabs.com/blog/agents-md</link><guid isPermaLink="true">https://elrootlabs.com/blog/agents-md</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our AGENTS.md runs to 395 lines, about twice what Claude Code&apos;s docs suggest. What goes in it, what we keep out, and the bug behind each rule.</description></item>
<item><title>AI agent governance for a small team: the rules our agents run under</title><link>https://elrootlabs.com/blog/ai-agent-governance</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-governance</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>One owner, one control session, several AI agents. Our AI agent governance rules, who approves what, and the incident behind each rule.</description></item>
<item><title>AI agent observability: what to log on every tool call</title><link>https://elrootlabs.com/blog/ai-agent-observability</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-observability</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We read the log of a six-hour session with a coding agent: 886 tool calls. What it recorded, what we had to rebuild, and the record we now want for every call.</description></item>
<item><title>AI agent permissions: a least-privilege checklist for tools and APIs</title><link>https://elrootlabs.com/blog/ai-agent-permissions</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-permissions</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We mapped every permission our own writing agent holds. One limit was enforced in code. Most of the rest, including the key one, were lines in its prompt.</description></item>
<item><title>Is it safe to let an AI agent change production infrastructure? How we let ours do it</title><link>https://elrootlabs.com/blog/ai-agent-production-changes</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-production-changes</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>When an AI agent can safely change live infrastructure, shown through a real change an agent made to our hosting, and the four it chose not to make.</description></item>
<item><title>AI agent tool failures: the five kinds we hit in one week, and how to recover from each</title><link>https://elrootlabs.com/blog/ai-agent-tool-failures</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-tool-failures</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Seven tool failures from our own AI agents in one week, sorted into five kinds, with the recovery that fits each and the signal that tells them apart.</description></item>
<item><title>AI automation for small businesses: where to start without creating new risks</title><link>https://elrootlabs.com/blog/ai-automation-small-business</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-automation-small-business</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>AI automation for small business, one workflow at a time: how we automate our own small company with plain rules first and AI in one place, behind an approval.</description></item>
<item><title>AI trends for 2027, read from a year of search data</title><link>https://elrootlabs.com/blog/ai-trends-2027</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-trends-2027</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We put 98 technology searches on one scale for a year. Agents, coding and automation rose; LLM, RAG and MCP fell. What that suggests about AI trends in 2027.</description></item>
<item><title>Claude skills: what we found in the 53 our marketing agent can load</title><link>https://elrootlabs.com/blog/claude-skills</link><guid isPermaLink="true">https://elrootlabs.com/blog/claude-skills</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We audited the 53 Claude skills our marketing agent can load. 456 files, 3 runnable scripts and a lock file that pins nothing. The checklist we use now.</description></item>
<item><title>Computer vision label inspection, tested on 500 synthetic labels</title><link>https://elrootlabs.com/blog/computer-vision-label-inspection</link><guid isPermaLink="true">https://elrootlabs.com/blog/computer-vision-label-inspection</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We ran four computer vision label inspection checks on 500 synthetic labels. All caught missing and skewed labels; light caused the false rejects.</description></item>
<item><title>Context engineering: what we put around the prompt, measured in words</title><link>https://elrootlabs.com/blog/context-engineering</link><guid isPermaLink="true">https://elrootlabs.com/blog/context-engineering</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our daily writing agent could read about 107,000 words of files. Its playbook has it load about 20,000. What goes into its context, what stays out, and why.</description></item>
<item><title>Edge AI vs cloud AI: a cost calculation for 1,000 devices</title><link>https://elrootlabs.com/blog/edge-ai-vs-cloud-ai-cost</link><guid isPermaLink="true">https://elrootlabs.com/blog/edge-ai-vs-cloud-ai-cost</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>Edge AI vs cloud AI, worked out with November 2026 prices: cloud wins for occasional text requests, edge wins by far for a camera checking every few seconds.</description></item>
<item><title>Google AI Mode and your website: what to check, from Google&apos;s own rules</title><link>https://elrootlabs.com/blog/google-ai-mode-website</link><guid isPermaLink="true">https://elrootlabs.com/blog/google-ai-mode-website</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Google says AI Mode needs no special markup or files. We checked our own site against the rules it does set: four settings that can still keep you out.</description></item>
<item><title>Google Trends for keyword research: five traps we hit in our own data</title><link>https://elrootlabs.com/blog/google-trends-keyword-research</link><guid isPermaLink="true">https://elrootlabs.com/blog/google-trends-keyword-research</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>We used Google Trends twice in one week to pick blog topics. Five traps showed up in our data, from a sinking anchor term to queries that mean something else.</description></item>
<item><title>Hire AI developers or an AI development company? How to decide, and how to test them</title><link>https://elrootlabs.com/blog/hire-ai-developers</link><guid isPermaLink="true">https://elrootlabs.com/blog/hire-ai-developers</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Should you hire AI developers or work with an AI development company? Who carries which risk, the four ways to engage a firm, and four tests for any AI hire.</description></item>
<item><title>Humanoid robot vs cobot: what each costs per hour of work, and the payback</title><link>https://elrootlabs.com/blog/humanoid-robot-vs-cobot</link><guid isPermaLink="true">https://elrootlabs.com/blog/humanoid-robot-vs-cobot</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>robotics</category><description>A humanoid robot from $13.5K or a cobot arm at $9,500? We worked out cost per working hour from the makers&apos; spec sheets, and the payback at Indian labour costs.</description></item>
<item><title>Setting up IndexNow on a Next.js site, including scheduled posts</title><link>https://elrootlabs.com/blog/indexnow-nextjs</link><guid isPermaLink="true">https://elrootlabs.com/blog/indexnow-nextjs</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>How we wired IndexNow into a prebuilt Next.js site with scheduled posts: one key file, one workflow that waits for the rebuild, and two mistakes we fixed.</description></item>
<item><title>IoT SIM cards in India: what one device report costs in data, measured</title><link>https://elrootlabs.com/blog/iot-sim-card-india-data-per-device</link><guid isPermaLink="true">https://elrootlabs.com/blog/iot-sim-card-india-data-per-device</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>An 82-byte reading used 4,844 bytes over a fresh TLS connection. How to size an IoT SIM card in India from measured bytes, and DoT&apos;s 13-digit M2M rule.</description></item>
<item><title>Measuring LLM hallucination rate on your own documents</title><link>https://elrootlabs.com/blog/llm-hallucination-rate</link><guid isPermaLink="true">https://elrootlabs.com/blog/llm-hallucination-rate</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We asked two models questions our own passages could not answer. With a plain prompt, six of nine answers came from outside the passage. One line fixed it.</description></item>
<item><title>LLM structured output: no broken JSON in 80 calls, eight wrong answers</title><link>https://elrootlabs.com/blog/llm-structured-output</link><guid isPermaLink="true">https://elrootlabs.com/blog/llm-structured-output</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We ran 80 extraction calls with and without JSON schema enforcement. Every reply was valid JSON. All eight mistakes were in following the rules.</description></item>
<item><title>OTA firmware updates: designing for the update that fails, on an ESP32</title><link>https://elrootlabs.com/blog/ota-firmware-update-esp32</link><guid isPermaLink="true">https://elrootlabs.com/blog/ota-firmware-update-esp32</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>Design an OTA firmware update for the failure cases: flash budget for A/B slots on ESP32, the battery one update costs, and ESP-IDF&apos;s rollback rules.</description></item>
<item><title>Post-quantum cryptography: what ML-KEM and ML-DSA change for an app</title><link>https://elrootlabs.com/blog/post-quantum-cryptography-ml-kem</link><guid isPermaLink="true">https://elrootlabs.com/blog/post-quantum-cryptography-ml-kem</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>security</category><description>We timed ML-KEM and ML-DSA against X25519, Ed25519 and RSA, and checked which sites already use post-quantum TLS. Speed was fine. Size is the problem.</description></item>
<item><title>Prompt versioning: changing a prompt without breaking production</title><link>https://elrootlabs.com/blog/prompt-versioning</link><guid isPermaLink="true">https://elrootlabs.com/blog/prompt-versioning</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our writing agent&apos;s instructions changed five times in its first week. Each change shipped with the check that enforces it. What we version, and the gap.</description></item>
<item><title>A schema markup audit of our own site, before and after a rebuild</title><link>https://elrootlabs.com/blog/schema-markup-audit</link><guid isPermaLink="true">https://elrootlabs.com/blog/schema-markup-audit</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>We audited our own site&apos;s schema markup before and after a rebuild. Everything parsed both times; the move fixed one gap, kept one and added a new one.</description></item>
<item><title>SEO vs AEO vs GEO, plus AIO and SXO: what each changed on one of our pages</title><link>https://elrootlabs.com/blog/seo-vs-aeo-vs-geo</link><guid isPermaLink="true">https://elrootlabs.com/blog/seo-vs-aeo-vs-geo</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>SEO vs AEO vs GEO, plus AIO and SXO, shown on one real page. What each layer changed when we rewrote it, and what each can and cannot be measured with.</description></item>
<item><title>Website security checklist for small businesses: what to fix first</title><link>https://elrootlabs.com/blog/small-business-website-security-checklist</link><guid isPermaLink="true">https://elrootlabs.com/blog/small-business-website-security-checklist</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>security</category><description>A website security checklist for small businesses in the order to fix it, with the one-line test for each item and the results when we ran it on our own site.</description></item>
<item><title>Vibe coding to production: what our AI-built website needed before it held up</title><link>https://elrootlabs.com/blog/vibe-coding-to-production</link><guid isPermaLink="true">https://elrootlabs.com/blog/vibe-coding-to-production</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Most commits behind our website were co-written by an AI coding agent. These are the fixes it needed after the code worked, and the checks we run now.</description></item>
<item><title>Web5 vs Web3: what Web5 is, and why almost nobody searches for it</title><link>https://elrootlabs.com/blog/web5-vs-web3</link><guid isPermaLink="true">https://elrootlabs.com/blog/web5-vs-web3</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Web5 vs Web3, with our own search data: Web5 scored zero on Google Trends. Its ideas live on as identity standards, and those are the part worth building on.</description></item>
<item><title>Website maintenance checklist: what runs on our live site, and how often</title><link>https://elrootlabs.com/blog/website-maintenance-checklist</link><guid isPermaLink="true">https://elrootlabs.com/blog/website-maintenance-checklist</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>A website maintenance checklist from our own live site. What runs every 5 minutes, daily, weekly and before each release, and two schedules that ran hours late.</description></item>
<item><title>Website migration checklist: five things that broke when we moved our site</title><link>https://elrootlabs.com/blog/website-migration-checklist</link><guid isPermaLink="true">https://elrootlabs.com/blog/website-migration-checklist</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>We kept every URL when we moved our site to a new stack, and five things broke anyway. Our website migration checklist, built from what each one taught us.</description></item>
<item><title>How we took our homepage from 51 to 88 on Google&apos;s mobile speed test</title><link>https://elrootlabs.com/blog/website-speed-core-web-vitals</link><guid isPermaLink="true">https://elrootlabs.com/blog/website-speed-core-web-vitals</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Five fixes that improved our Core Web Vitals in one day, each with its before and after number, in the order to check them on your own slow website.</description></item>
<item><title>How we built an AI blog pipeline, and what it checks before a post ships</title><link>https://elrootlabs.com/blog/ai-blog-pipeline</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-blog-pipeline</guid><pubDate>Tue, 17 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our AI content pipeline writes only topics a person approved, then a 330-line script with no model in it rejects tells, unsourced numbers and unknown sources.</description></item>
<item><title>RAG chunking strategies, tested on the same document set</title><link>https://elrootlabs.com/blog/rag-chunking-strategies</link><guid isPermaLink="true">https://elrootlabs.com/blog/rag-chunking-strategies</guid><pubDate>Tue, 17 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We ran eight RAG chunking strategies over our own blog with the same 24 questions. Labelling each chunk with its title and heading beat making chunks bigger.</description></item>
<item><title>Designing the approval step for AI agents: what the reviewer needs to see</title><link>https://elrootlabs.com/blog/agent-approval-step-design</link><guid isPermaLink="true">https://elrootlabs.com/blog/agent-approval-step-design</guid><pubDate>Sat, 14 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>What a human approval step for an AI agent has to show the reviewer, and the four ways the step fails quietly while still looking like oversight.</description></item>
<item><title>Denial of wallet: AI agent cost control before the bill arrives</title><link>https://elrootlabs.com/blog/ai-agent-cost-control</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-cost-control</guid><pubDate>Sat, 14 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>What makes an AI agent run up a bill, from runaway loops to denial of wallet attacks, and the limits that stop it before the invoice does.</description></item>
<item><title>How to build an evaluation set for an AI agent from your own failure cases</title><link>https://elrootlabs.com/blog/agent-evaluation-test-set</link><guid isPermaLink="true">https://elrootlabs.com/blog/agent-evaluation-test-set</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Build an AI agent evaluation set from the failures that cost you trust, time or money. A worked example turns one incident into five cases and a release rule.</description></item>
<item><title>AI context platforms and the engineering problem underneath them</title><link>https://elrootlabs.com/blog/ai-context-platform</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-context-platform</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Context is a finite resource with diminishing returns. What an AI context platform has to solve, and the four techniques doing that work today.</description></item>
<item><title>ChatGPT search optimization: how to get your business cited in ChatGPT and Google AI answers</title><link>https://elrootlabs.com/blog/ai-search-optimization-chatgpt</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-search-optimization-chatgpt</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>ChatGPT search optimization comes down to the crawler you allow, snippet eligibility and honest markup. With a week of our own crawler data.</description></item>
<item><title>Cloud vs on-premise for small businesses in 2027</title><link>https://elrootlabs.com/blog/cloud-vs-on-premise-2027</link><guid isPermaLink="true">https://elrootlabs.com/blog/cloud-vs-on-premise-2027</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Decide per system, not per company. What must keep working when the line drops, what the provider will never do for you, and what moving back actually costs.</description></item>
<item><title>Computer vision quality inspection: where it pays in manufacturing</title><link>https://elrootlabs.com/blog/computer-vision-quality-inspection</link><guid isPermaLink="true">https://elrootlabs.com/blog/computer-vision-quality-inspection</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Computer vision quality inspection pays when defects are visible and costly. How to choose a method when you have few or no photos of real defects.</description></item>
<item><title>Custom software vs off-the-shelf: what building our own website taught us</title><link>https://elrootlabs.com/blog/custom-software-vs-off-the-shelf</link><guid isPermaLink="true">https://elrootlabs.com/blog/custom-software-vs-off-the-shelf</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Custom software vs off-the-shelf, from a company that built its own website and CMS: what it cost, what it bought, and a test for your own decision.</description></item>
<item><title>Drone development in India: the rules to design for from day one</title><link>https://elrootlabs.com/blog/drone-development-india</link><guid isPermaLink="true">https://elrootlabs.com/blog/drone-development-india</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>robotics</category><description>Drone development in India starts with the Drone Rules: weight class, type certificate, Digital Sky registration, airspace zones and radio approval.</description></item>
<item><title>Fine-tuning vs RAG: how to choose, and when a better prompt wins</title><link>https://elrootlabs.com/blog/fine-tuning-vs-rag</link><guid isPermaLink="true">https://elrootlabs.com/blog/fine-tuning-vs-rag</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>RAG fixes missing facts, fine-tuning fixes behaviour. We tested retrieval on our own blog and found where it really fails: the right page, the wrong paragraph.</description></item>
<item><title>Fixed price vs time and materials: choosing a software contract</title><link>https://elrootlabs.com/blog/fixed-price-vs-time-and-materials</link><guid isPermaLink="true">https://elrootlabs.com/blog/fixed-price-vs-time-and-materials</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Fixed price suits a scope you can write down today. Time and materials suits work you will discover as you build. How to choose, and what moves the cost.</description></item>
<item><title>Guardian agents: what runtime oversight of AI agents actually requires</title><link>https://elrootlabs.com/blog/guardian-agents-runtime-oversight</link><guid isPermaLink="true">https://elrootlabs.com/blog/guardian-agents-runtime-oversight</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>What guardian agents are, why pre-deployment checks miss the failures that matter, and what the research says a supervisory layer has to watch at runtime.</description></item>
<item><title>IoT prototype development in India: from dev kit to an approved device</title><link>https://elrootlabs.com/blog/iot-prototype-development-india</link><guid isPermaLink="true">https://elrootlabs.com/blog/iot-prototype-development-india</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>IoT prototype development in India in seven stages, with the WPC and BIS approvals built in from the first board so your design does not block the launch.</description></item>
<item><title>MCP server security: what the protocol does not check for you</title><link>https://elrootlabs.com/blog/mcp-server-security</link><guid isPermaLink="true">https://elrootlabs.com/blog/mcp-server-security</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>MCP server security starts from an uncomfortable fact. The protocol does not verify which agent is calling, so identity and scope are yours to add.</description></item>
<item><title>Why multi-agent systems fail: 14 failure modes and what they share</title><link>https://elrootlabs.com/blog/multi-agent-system-failures</link><guid isPermaLink="true">https://elrootlabs.com/blog/multi-agent-system-failures</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Multi-agent system failures cluster into three categories and fourteen modes. What the research found in 1,600 traces, and why more agents rarely help.</description></item>
<item><title>MVP development: from idea to first users, with our own pipeline as the example</title><link>https://elrootlabs.com/blog/mvp-development</link><guid isPermaLink="true">https://elrootlabs.com/blog/mvp-development</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>MVP development in four phases, shown on a real example: the blog pipeline we built, what version one left out, and what real use forced us to change.</description></item>
<item><title>PCB layout mistakes to catch before you order the first board</title><link>https://elrootlabs.com/blog/pcb-layout-mistakes</link><guid isPermaLink="true">https://elrootlabs.com/blog/pcb-layout-mistakes</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>Eight PCB layout mistakes that sink a first IoT board, each paired with the chip maker&apos;s own rule, so you can check the design before you pay for a run.</description></item>
<item><title>Running an LLM offline: what it takes and when it is worth it</title><link>https://elrootlabs.com/blog/run-llm-offline</link><guid isPermaLink="true">https://elrootlabs.com/blog/run-llm-offline</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>An offline LLM keeps data on the device and works without a network, at the cost of a smaller model. Five decisions settle whether it fits your app.</description></item>
<item><title>Sovereign AI: where your model runs and who can reach it</title><link>https://elrootlabs.com/blog/sovereign-ai</link><guid isPermaLink="true">https://elrootlabs.com/blog/sovereign-ai</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Sovereign AI covers where a model runs, who operates the machines, and who can be compelled to grant access. What the deployment models actually differ on.</description></item>
<item><title>Who owns the code when you hire a software company in India?</title><link>https://elrootlabs.com/blog/who-owns-the-code-india</link><guid isPermaLink="true">https://elrootlabs.com/blog/who-owns-the-code-india</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Under India&apos;s Copyright Act, a software company owns the code it writes for you unless your software development agreement assigns it in writing.</description></item>
<item><title>20 prompt-injection test cases to run against your AI agent before launch</title><link>https://elrootlabs.com/blog/prompt-injection-test-cases</link><guid isPermaLink="true">https://elrootlabs.com/blog/prompt-injection-test-cases</guid><pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>A copyable prompt injection testing set for agents that read documents, email and web pages, grouped by where hostile text hides, each with its safe behaviour.</description></item>
<item><title>What safeguards should control an AI agent in production? The eight we run on our own</title><link>https://elrootlabs.com/blog/production-ai-agent-safety</link><guid isPermaLink="true">https://elrootlabs.com/blog/production-ai-agent-safety</guid><pubDate>Sat, 31 Oct 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>The eight safeguards an AI agent needs in production, the order to add them, who signs off, and how each one works on the agent that runs our own publishing.</description></item>
<item><title>Local-first software is an operational decision</title><link>https://elrootlabs.com/blog/local-first-field-software</link><guid isPermaLink="true">https://elrootlabs.com/blog/local-first-field-software</guid><pubDate>Sun, 06 Sep 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>When connectivity is uncertain, offline behaviour is not a fallback screen. It becomes part of the product’s core data model.</description></item>
<item><title>Build school software around the institution, not the dashboard</title><link>https://elrootlabs.com/blog/school-software-institution-first</link><guid isPermaLink="true">https://elrootlabs.com/blog/school-software-institution-first</guid><pubDate>Sun, 06 Sep 2026 00:00:00 GMT</pubDate><category>education-technology</category><description>Education systems succeed when they respect the daily work of teachers, administrators, students and families—not when they maximise feature count.</description></item>
</channel>
</rss>
