Articles, one at a time.
Every piece here was commissioned, drafted, reviewed in public, and merged. No content mills, no auto-published slop.
Codex Goal Mode: Run Long-Horizon AI Coding Tasks for Hours or Days
Codex Goal Mode is now stable. Learn how /goal turns Codex into a persistent coding agent that runs for hours or days without re-prompting.
Read →
Gemini 3.5 Flash API: Google I/O 2026 Developer Guide
Gemini 3.5 Flash launched at Google I/O 2026 with 76.2% Terminal-Bench 2.1 and 1M context. Effloow Lab verified model ID, pricing, benchmarks, and migration path.
Read →
Genkit 2.0 GA: Firebase AI, MCP, and Cloud Run
Effloow Lab ran a local Genkit MCP PoC: typed tools, MCP discovery, Cloud Run deployment path, and production limits for Firebase AI apps.
Read →
Mnemonic Sovereignty: LLM Agent Memory Security Attacks Explained
arXiv:2604.16548 surveys six attack phases targeting LLM agent long-term memory. Effloow Lab ran a Python PoC demonstrating memory poisoning and a trust-gate mitigation.
Read →
Mastra 1.0: TypeScript-First AI Agent Framework Developer Guide
Mastra 1.0 is a TypeScript agent framework with native MCP and A2A support. Effloow Lab inspected @mastra/core 1.37.0 — here's what you need to know.
Read →
AgentAtlas: LLM Agent Benchmarks Need More Than Accuracy
AgentAtlas introduces a 6-state control-decision taxonomy and 9-category failure taxonomy to expose what accuracy columns hide about LLM agents.
Read →
Claude MCP Tunnels: Private MCP Access for Agents
A developer guide to Claude MCP Tunnels, with local sandbox evidence for hostname routing, deny-by-default access, and limits.
Read →
NVIDIA Nemotron 3 Nano Omni: Multimodal Agent Dev Guide
NVIDIA Nemotron 3 Nano Omni: 30B-A3B MoE, 256K context, audio/video/image/text unified. Deploy with vLLM, NIM, or use free via OpenRouter.
Read →
Google Managed Agents: Gemini API Runtime Guide 2026
A developer guide to Google Managed Agents in the Gemini API, with local sandbox evidence for agent package and policy design.
Read →
Reward Hacking in LLM Agents: What the RHB Benchmark Reveals
RHB benchmark (arXiv:2605.02964) shows RL-trained agents exploit tool-use environments. Learn what triggers reward hacking and how to harden your agent setup.
Read →
SubQ: Sub-Quadratic LLM with 12M Token Context — 2026 Guide
SubQ is the first sub-quadratic LLM with a 12M token context window. Learn how SSA architecture works, what it costs, and how to get API access.
Read →
AI Pair Programming to Autonomous Teams: A 2026 Workflow
A practical 2026 workflow for moving from AI pair programming to delegated coding-agent teams with review gates and sandbox evidence.
Read →