Saturday, November 7, 2026 270 picks · 133 builders credited Subscribe
AgentRags.

The agentic rag, wrung out.

The rag for

Monday, October 19, 2026

The day in one breath

OpenClaw 2026.10.5 stable release

It stands out because it is a large, practical reliability upgrade (self-healing updates + restart-safe conversations) combined with new user-facing surface area (X mentions channel with cost controls, richer voice calls, live model discovery). For a project that runs as a long-lived local/server agent, reliability of the Gateway and conversation continuity is the highest-leverage change. The accompanying community posts and official release notes make it the clearest single story that active OpenClaw users will care about today.

Releases

OpenClaw 2026.10.5: Self-healing Updates, Restart-safe Conversations, X Channel & Voice Calls

Major stable release focused on reliability. Updates now finish and recover themselves across platforms. Conversations, subagent work, and deferred tasks resume cleanly after Gateway restarts. Gateway is faster and lighter (session/history reads moved to workers). New X (Twitter) mentions channel with streaming and cost limits. Voice calls gain briefs, reports, live steering, and voicemail detection. Smarter model pickers surface new Claude/ChatGPT/Ollama models immediately. ClawHub browsing added to Gateway. Tool approval hooks tightened so plugins cannot silently rewrite approved arguments.

Guides & Workflows

Five Hermes Features Worth Knowing + Trust: Untrusted for MCP

witcheer highlighted five features (three already in v0.22.2, two incoming): agent-assisted plugin/skill install, Desktop tour, auto-discovery of installed apps for plugins, live countdown for repeating tasks, and confirmation before discarding unfinished code. Separate post covers `trust: untrusted` in config.yaml so MCP tools marked non-read-only require explicit /approve.

Multi-agent & Memory

Paperclip — Org Chart, Budgets, and Heartbeats for Teams of Agents (Supports OpenClaw & Hermes)

Open-source control plane that turns multiple agents into a managed company. Org charts, per-agent budgets that pause work when exhausted, task ownership, review gates, and heartbeats. Explicitly supports OpenClaw, Hermes, Claude Code, Codex, Cursor, and others via adapters. One-line onboard: `npx paperclipai onboard --yes`. Recent releases improved agent file persistence and Grok Build support.

Releases

Hermes Agent to Become a DIY Game Dev Studio

Teknium posted that he is turning Hermes Agent into a DIY game development studio. Follow-up thread details needed plugins: 3D model generation, animation, sound FX, Unity/Godot integration (Unreal already available), Blender return, sprite sheets, music, character voices, playtester agents, and one-click export.

Never miss a night

Every pick, delivered to your reader.

One feed for everything, or one per agent harness you run. Rebuilt every night, with full credit on every item.