Site-wide changelog: tools, reviews, posts, protocols, and major site changes — all in one feed.https://labgrimoire.com/https://labgrimoire.com/blog/tmux-bridge-mcp-multi-agent-terminal/post:tmux-bridge-mcp-multi-agent-terminal:enRunning multiple AI coding agents but still copy-pasting between terminals by hand? tmux-bridge-mcp lets agents read and write each other's terminals over MCP, with install steps, supply-chain security review notes, and multi-agent terminal tuning tipsWed, 22 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/cloudflare-tunnel-ai-agent-second-mac/post:cloudflare-tunnel-ai-agent-second-mac:enUse Cloudflare Tunnel and Zero Trust to open a multi-device AI Agent channel, secure SSH into an idle iMac as a LINE Bot extension, with full setup steps and failure notesTue, 21 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/apple-silicon-mlx-lora-33-seconds/post:apple-silicon-mlx-lora-33-seconds:enRun LoRA local fine-tuning of Gemma-4 E4B on an M4 Pro with mlx-tune. In 33 seconds the model learns your JSON schema. A full field report of Apple Silicon local fine-tuning.Mon, 20 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/windows-qlora-brand-voice-finetune/post:windows-qlora-brand-voice-finetune:enFine-tune Qwen3.5 with QLoRA on an RTX 4060 Ti using 76 ChatML samples. A full walkthrough from environment setup through LoRA merge, GGUF conversion, and Ollama deploy, plus five hard rules for Windows CUDA fine-tuning.Sun, 19 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/local-llm-line-customer-service-bot/post:local-llm-line-customer-service-bot:enA hands-on guide to deploying a LINE Bot customer service agent on Windows GPU with Ollama, Qwen3.5, and Hermes Gateway, including six real production pitfalls and fixesSat, 18 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/dual-brain-kanban/post:dual-brain-kanban:enHow do you let the same Kanban board be programmatically managed by any AI Agent while rendering beautifully in real-time via Claude Desktop's Live artifacts, even letting you edit cards straight from the UI? Breaking down the three-layer co-governance architecture of data, MCP bridge, and presentation, with a full walkthrough of my actual Morning War Room implementation and security red lines.Fri, 17 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/2026-ai-agent-toolchain-comparison/post:2026-ai-agent-toolchain-comparison:enHorizontal comparison of five mainstream AI Agent tools: memory persistence, Hook extensibility, Skill routing, and cross-platform integration, with scenario-based recommendations.Thu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/changelog/site:fate-subdomain• The Fate & Divination culture tool now lives at fate.labgrimoire.com, kept separate from the science toolset; the tools directory card links straight there • Old /tools/fate links redirect to the new subdomain, so existing bookmarks keep workingThu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/changelog/site:fate-ziwei-patterns• Birth charts in the Fate & Divination tool now highlight the four transformations (化祿/化權/化科/化忌) with colour-coded markers on the affected stars • The tool also spots traditional 格局 (star patterns) in your chart and explains each one in plain, for-entertainment language with no fortune-telling claimsThu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/changelog/site:tool-cards-full-page• Selecting a tool from the directory now opens it directly in full-page analysis mode, so you land in the maximum working area straight away • The separate Full Page button on each card was removed because the whole card now leads there; use the in-tool Exit full page control to return to the embedded viewThu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/tools/datawash/tool:datawash:v1.1• Layout refresh: detection findings now use divider-separated collapsible groups instead of nested cards • Consistent corner radius and spacing rhythm aligned with the site design tokens • Improved mobile density with a dedicated small-screen breakpointThu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/tools/fate/tool:fate:v1.1• Repositioned from entertainment framing to a cultural-studies perspective across the interface and readings; honest no-prediction boundaries unchanged • Results now use tab switching (Bazi ↔ Zi Wei, Meihua ↔ Xiao Liu Ren) instead of one long scroll — both charts are still computed at once • Reading paragraphs gain semantic labels (career / caution / strengths / advice) so each line's role is clear at a glance • Collapsible glossary added to each system explaining core terms in plain language • Corner radius and spacing aligned with the site design tokensThu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/tools/fate/tool:fate:v1.2• New paper-white editorial look with a cinnabar accent and serif headings, scoped to this tool and separate from the site's dark theme • Birth-chart and Question zones are now top-level tabs; form state is preserved when switchingThu, 16 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/ai-agent-7-fatal-mistakes/post:ai-agent-7-fatal-mistakes:enSeven real-world AI Agent disaster stories from six months of use: from rm -rf burns and API Key leaks to total cross-session amnesia, every mistake became a system safeguard.Wed, 15 Jul 2026 00:00:00 GMThttps://labgrimoire.com/changelog/site:polar-billing• Lab Pro is now available for purchase at $19/month or $190/year (save 17% annually) • Lab Pro raises the daily download and export limit from 5 to 50, the only paid benefit, and every analysis tool and feature remains free for all users • Invoices and tax handling are generated automatically with each payment • Subscriptions can be managed or cancelled anytime from a self-service billing portalWed, 15 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/complete-agent-workflow-deployment/post:complete-agent-workflow-deployment:enA hands-on guide to building an AI Agent system: from directory layout and boot config to three-layer memory architecture and hook-based guardrails, seven steps to a working Agent workflow.Tue, 14 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/complete-agent-workflow-blueprint/post:complete-agent-workflow-blueprint:enDesign a complete AI Agent system architecture from the ground up, covering six layers: identity, memory, routing, gating, cross-platform, and safety. Understand what each layer solves and how they interconnect.Mon, 13 Jul 2026 00:00:00 GMThttps://labgrimoire.com/changelog/site:llms-v2-2• The AI discovery guide now includes the Stroop Test • A new NAD+ evidence guide leads with the 113-study PRISMA systematic review and distinguishes biological or biomarker activity from unestablished clinical benefits • Researchers can now find all four practical AI workflow and configuration resources directly from the discovery guideMon, 13 Jul 2026 00:00:00 GMThttps://labgrimoire.com/changelog/site:seo-crawl-health• Language alternates now point only to pages that actually exist, across desktop and mobile navigation • Account and billing routes are excluded from search indexing and public sitemaps • Research resources no longer send readers or crawlers to unpublished guidesMon, 13 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/agent-safety-baseline-design/post:agent-safety-baseline-design:enAI Agent automation brings efficiency but also risk. This article shares 10 battle-tested safety baseline rules with Hook gatekeeper implementation concepts, helping you build an AI workflow that won't self-destruct.Sun, 12 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/context-dependent-geroprotectors-sex-stage-strain/post:context-dependent-geroprotectors-sex-stage-strain:enResveratrol, rapamycin, and diet all claim to fight aging, so why do the effects vary so much? Studies in 2026 show anti-aging interventions succeed or fail depending on sex, life stage, and genetic background, pushing geroscience from a lifespan race toward hallmark-informed prioritization.Sat, 11 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/voidnotch-open-source/post:voidnotch-open-source:enMIT open-source macOS notch dashboard for system health and AI coding-provider quotas. Local-first: clone, build, and run.Sat, 11 Jul 2026 00:00:00 GMThttps://labgrimoire.com/spellbook/voxcpm/spell:voxcpm:enVoxCPM2 is a 2B-parameter open-source TTS engine supporting 30 languages, voice design, and voice cloning. Best suited for developers and educators who want local, multilingual speech synthesis without recurring API fees. Requires a CUDA GPU for practical speed.Thu, 09 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/lora-finetune-speak-your-language/post:lora-finetune-speak-your-language:enLoRA and QLoRA let you fine-tune a large language model with just a few dozen examples on your own computer, no cloud GPU rental, no uploading sensitive data. A real biotech lab case shows why this technique changes how local AI feels to use.Wed, 08 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/yif1a-golgi-mtorc1-senescence-switch/post:yif1a-golgi-mtorc1-senescence-switch:enmTORC1 is the most famous molecular switch in aging research, yet how it gets turned on at the Golgi has stayed murky. A 2026 study shows the Golgi protein YIF1A recruits RNF126 to degrade G3BP1/2, releasing the brake on mTORC1; cells with YIF1A knocked down resist induced senescence, and worm experiments reveal the pathway tunes lifespan in both directions.Wed, 08 Jul 2026 00:00:00 GMThttps://labgrimoire.com/spellbook/tronclass/spell:tronclass:enA closed-source SaaS learning management system from a Taiwanese vendor. Its AIR module auto-generates lecture highlights, subtitles, quizzes, and Q&A. Best suited for university instructors who need a turnkey Chinese-first LMS with proven campus adoption. Limited by opaque pricing and zero API access for automation.Wed, 08 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/aging-brain-inter-organ-immune-glucose-vascular/post:aging-brain-inter-organ-immune-glucose-vascular:enFour 2026 studies converge on one conclusion: brain aging is not the brain's problem alone. Peripheral CD8⁺ EMRA T-cell immunosenescence, plasma glucose with Mendelian Randomization causal evidence, and arterial wall cellular heterogeneity collapse are converging on the aging brain — while intracranial microglia and oligodendrocytes inflammage in near-perfect synchrony (Pearson r≈0.98). Understanding this multi-source siege reframes what prevention might realistically look like.Tue, 07 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/google-agents-cli-progressive-disclosure-architecture-analysis/post:google-agents-cli-progressive-disclosure-architecture-analysis:enGoogle agents-cli uses a three-tier SkillToolset progressive disclosure architecture that cuts baseline context token consumption by about 90% in static scenarios, though real-world workloads may narrow that to 37%. This piece dissects the token-efficiency claim, the vendor lock-in risk, and the fragility of description-based routing, extracting lessons useful for any independent agent system.Tue, 07 Jul 2026 00:00:00 GMThttps://labgrimoire.com/spellbook/taste-skill/spell:taste-skill:enA portable SKILL.md framework that injects design taste into AI coding agents (Claude Code, Cursor, Codex). Three tunable knobs control layout variance, motion intensity, and visual density. Zero-friction install, framework-agnostic. Worth watching, but v2 is still experimental and results depend heavily on LLM quality.Tue, 07 Jul 2026 00:00:00 GMThttps://labgrimoire.com/blog/local-llm-benchmark-vs-cloud-tier/post:local-llm-benchmark-vs-cloud-tier:enI benchmarked five 4-bit local models on a single M4 Pro (48GB) with oMLX across eight benchmarks, mapped the speed/quality sweet spot, and calibrated which cloud generation they match. Plus a field report on three speculative-decoding traps (MTP / dflash / specprefill) and the cache trade-off that decides them.Mon, 06 Jul 2026 00:00:00 GMT