<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Agent Reliability Lab materials</title><description>Operator notes on verification, reliability, local models, automation, and the web.</description><link>https://ss-global-group.com/</link><item><title>Why AI agents game their own tests, and how to isolate verification</title><link>https://ss-global-group.com/materials/ai-agents-game-their-tests/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-agents-game-their-tests/</guid><description>An agent that writes the test will pass it. Isolate the falsifier, allow BLOCKED, and stop treating a green self-report as a release signal.</description><pubDate>Sat, 05 Sep 2026 00:00:00 GMT</pubDate></item><item><title>A macro dashboard built by a multi-agent ensemble on DuckDB</title><link>https://ss-global-group.com/materials/multi-agent-macro-dashboard/</link><guid isPermaLink="true">https://ss-global-group.com/materials/multi-agent-macro-dashboard/</guid><description>Beat rate is a first derivative. A DuckDB panel plus a three-family ensemble reads acceleration — with a falsifier on every scenario.</description><pubDate>Sat, 05 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Multi-dimensional portfolio stress-testing against liquidity shocks</title><link>https://ss-global-group.com/materials/portfolio-stress-testing/</link><guid isPermaLink="true">https://ss-global-group.com/materials/portfolio-stress-testing/</guid><description>VaR is calm-weather maths. A 12-by-6 shock matrix and a 60 percent margin ceiling show where a book dies in a liquidity event.</description><pubDate>Sat, 05 Sep 2026 00:00:00 GMT</pubDate></item><item><title>PWA and Telegram Mini Apps: mobile without the app stores</title><link>https://ss-global-group.com/materials/pwa-and-telegram-mini-apps/</link><guid isPermaLink="true">https://ss-global-group.com/materials/pwa-and-telegram-mini-apps/</guid><description>A catalog, a booking flow, or a request form rarely needs an app-store cycle. One codebase can live as a PWA and a Telegram Mini App.</description><pubDate>Sat, 05 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Silent no-ops and fail-plausible: when the system reports success and does nothing</title><link>https://ss-global-group.com/materials/silent-noop-fail-plausible/</link><guid isPermaLink="true">https://ss-global-group.com/materials/silent-noop-fail-plausible/</guid><description>Green dashboards and exit 0 are claims. Silent no-ops freeze data; fail-plausible agents invent a success report. How operators catch both.</description><pubDate>Sat, 05 Sep 2026 00:00:00 GMT</pubDate></item><item><title>How to verify an LLM answer: a verification layer that catches hallucinations</title><link>https://ss-global-group.com/materials/how-to-verify-llm-answers/</link><guid isPermaLink="true">https://ss-global-group.com/materials/how-to-verify-llm-answers/</guid><description>Fluent is not true. Grounding, citations, an isolated judge, code for numbers, and a human on the high-stakes path — before a fabricated fact ships.</description><pubDate>Fri, 04 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Why leads from your website never arrive</title><link>https://ss-global-group.com/materials/why-website-leads-never-arrive/</link><guid isPermaLink="true">https://ss-global-group.com/materials/why-website-leads-never-arrive/</guid><description>Ads are live, analytics looks healthy, the CRM is empty. The leak is usually the delivery chain from the form to the desk — not the traffic.</description><pubDate>Fri, 04 Sep 2026 00:00:00 GMT</pubDate></item><item><title>The chatbot went silent: debugging failures that produce no error</title><link>https://ss-global-group.com/materials/chatbot-fails-without-errors/</link><guid isPermaLink="true">https://ss-global-group.com/materials/chatbot-fails-without-errors/</guid><description>Process up, logs clean, bot silent. Messenger bots fail without an exception. An external ping notices before a customer has to.</description><pubDate>Mon, 31 Aug 2026 00:00:00 GMT</pubDate></item><item><title>The price of privacy: local models vs API, in numbers</title><link>https://ss-global-group.com/materials/local-models-vs-api-cost/</link><guid isPermaLink="true">https://ss-global-group.com/materials/local-models-vs-api-cost/</guid><description>A 24 GB card, a rental GPU, or an API: the invoice is not the decision. Data that cannot leave the room is.</description><pubDate>Sun, 23 Aug 2026 00:00:00 GMT</pubDate></item><item><title>What a local model can actually do: 75 business tasks, measured</title><link>https://ss-global-group.com/materials/what-local-models-can-do/</link><guid isPermaLink="true">https://ss-global-group.com/materials/what-local-models-can-do/</guid><description>Two open-weight models on an office GPU, 75 office tasks, a code grader. Where a local assistant holds, where it invents, and what one rule changes.</description><pubDate>Sun, 23 Aug 2026 00:00:00 GMT</pubDate></item><item><title>An appointment bot that talks to the CRM in real time</title><link>https://ss-global-group.com/materials/appointment-bot-crm-integration/</link><guid isPermaLink="true">https://ss-global-group.com/materials/appointment-bot-crm-integration/</guid><description>A lead form collects a phone. An appointment bot reads live slots, locks one, and writes the CRM before anyone calls back.</description><pubDate>Fri, 21 Aug 2026 00:00:00 GMT</pubDate></item><item><title>What actually moves rankings</title><link>https://ss-global-group.com/materials/what-moves-search-rankings/</link><guid isPermaLink="true">https://ss-global-group.com/materials/what-moves-search-rankings/</guid><description>There is no official table of 200 weighted factors. Google ranks a mix of relevance, honesty, and technical fitness — and punishes bought visibility.</description><pubDate>Fri, 21 Aug 2026 00:00:00 GMT</pubDate></item><item><title>Usability audit: what it finds and what it&apos;s worth</title><link>https://ss-global-group.com/materials/usability-audit/</link><guid isPermaLink="true">https://ss-global-group.com/materials/usability-audit/</guid><description>A usability audit is a job-completion check, not a taste note. Formats are slice, funnel, or fixes — not a public price list.</description><pubDate>Thu, 20 Aug 2026 00:00:00 GMT</pubDate></item><item><title>AI Agent Drift: The Five Failure Layers, Diagnosed</title><link>https://ss-global-group.com/materials/ai-agent-drift/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-agent-drift/</guid><description>AI agent drift isn&apos;t one bug — it&apos;s five distinct failures in five harness layers. We run 10+ agents in production. Find the layer first, then fix it.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate></item><item><title>AI Agent Forgets Context Between Sessions — Why &amp; the Fix</title><link>https://ss-global-group.com/materials/ai-agent-forgets-context/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-agent-forgets-context/</guid><description>Your AI agent forgets everything between sessions because working memory dies at the session boundary. The three-case test for what to write down — and what not to.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate></item><item><title>AI Agent Infinite Loop: Fix the Death Spiral — Diagnose It</title><link>https://ss-global-group.com/materials/ai-agent-infinite-loop/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-agent-infinite-loop/</guid><description>Cursor or Claude Code agent stuck in a retry loop, editing the same file, burning tokens going in circles? Three signs, the math behind it, and a 4-rule fix.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate></item><item><title>AI Agent Says &quot;Done&quot; But Nothing Changed — Why &amp; How to Verify</title><link>https://ss-global-group.com/materials/ai-agent-says-done/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-agent-says-done/</guid><description>Your AI agent claims the task is done, but the file is unchanged and the test is red. Here&apos;s why agents over-claim — and how to measure the gap in 10 minutes.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate></item><item><title>How Big Should a .cursorrules / CLAUDE.md File Be?</title><link>https://ss-global-group.com/materials/cursorrules-size/</link><guid isPermaLink="true">https://ss-global-group.com/materials/cursorrules-size/</guid><description>How big should .cursorrules, CLAUDE.md, or AGENTS.md be? Two numbers to measure today, the lost-in-the-middle myth tested, and a refactor that lifted task success 45%→72%.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate></item><item><title>The Verification Gate Your AI Agent Is Skipping</title><link>https://ss-global-group.com/materials/verification-gate-ai-agent/</link><guid isPermaLink="true">https://ss-global-group.com/materials/verification-gate-ai-agent/</guid><description>Stop your AI agent claiming &apos;done&apos; on a broken build. A 30-line local check.py — lint, types, tests, smoke — that exits non-zero so &apos;done&apos; has to be true.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate></item><item><title>Why an AI agent works one day and not the next: drift at the harness level</title><link>https://ss-global-group.com/materials/ai-agent-inconsistent-runs/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-agent-inconsistent-runs/</guid><description>An agent that worked yesterday can quietly do the wrong thing today. That is harness-level drift: the world moved, the instruction did not.</description><pubDate>Tue, 30 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Website chatbot: template or custom</title><link>https://ss-global-group.com/materials/website-chatbot-template-or-custom/</link><guid isPermaLink="true">https://ss-global-group.com/materials/website-chatbot-template-or-custom/</guid><description>A widget covers FAQ and first-touch capture. A custom bot is the move when the catalog, the stock, or the slot grid has to be true.</description><pubDate>Fri, 26 Jun 2026 00:00:00 GMT</pubDate></item><item><title>AI, agentic system, autonomous agent: three levels of automation for an owner</title><link>https://ss-global-group.com/materials/ai-vs-agents/</link><guid isPermaLink="true">https://ss-global-group.com/materials/ai-vs-agents/</guid><description>An owner who asks for AI may mean a chatbot, a scripted fleet, or a digital employee. Three levels, and which one to start with.</description><pubDate>Sun, 31 May 2026 00:00:00 GMT</pubDate></item></channel></rss>