The agentic rag, wrung out.
The rag for
Hermes Agent v0.22.7 with opt-in lead bot routing for group chats. It is the most concrete, immediately usable multi-agent improvement released in the last 24 hours. The problem it solves (every bot waking on unaddressed messages) is universal among people running rooms, the solution is opt-in and measurable, and it shipped with supporting features (cron-into-room, live bot state). Combined with the Tokenjuice plugin and ongoing Cloud polish, it continues the pattern of Hermes shipping small, high-leverage controls rather than just new model wrappers. OpenClaw’s Particle-1 efficiency claim is strong but secondary today.
Nous Research released Hermes Agent v0.22.7 (Nov 3). The standout feature is opt-in lead routing in Desktop Group Chats. Previously, any message that named nobody woke every bot in the room. Now a designated lead bot can read the message and hand it to the right agent. Cron jobs can also post directly into rooms and wake only @mentioned bots. Full curated notes are deferred to v0.23.0; this is a patch roll-up of ~174 PRs.
A detailed write-up of discovering runaway heartbeat costs (~$2/day on useless polls, 62/63 doing nothing, context growing across polls). Includes a ready-to-paste read-only audit prompt for Claude Code that inventories cron, heartbeats, launchd/crontab, and standalone scripts; ranks cost; checks for context contamination and silent “OK” replies; and outputs a markdown report with ranked fixes. New rules suggested: never change model in place, log healthy checks, isolate sessions.
Clear guide covering four migration routes into Hermes Cloud: (1) `hermes migrate` from OpenClaw / Claude Code / Codex / Cursor / Grok Bot / OpenBot, (2) dashboard backup/restore from an existing local Hermes, (3) Desktop app side-by-side view of local and Cloud agents, (4) `hermes profile install` from a git repo. Notes that restore replaces rather than merges, and memories/sessions stay per-machine.
OpenClaw shipped v2026.10.10 with Chat Details side panel for wide chats, Advisor course-correction suggestions, and split speech for Browser Talk. Concurrently, Particle-1 is available inside OpenClaw. @particle_api reports ~45% fewer output tokens than DeepSeek V4.1 Flash with comparable accuracy and zero data retention on standard inference. Update to v2026.10.10 to try it.
Tokenjuice landed in the Hermes plugin catalog. It compresses successful terminal output (e.g., a 700-line `ls` dropped from 44k characters to 774; `git status` from 670 to 223) while leaving failed commands untouched so full errors still reach the model. Runs fully locally; nothing is sent over the network. Requires Hermes ≥ 0.22.5. Install via the catalog or npm + `hermes plugins install`.
A practical map of nine open-source agent systems organized by style: bot rosters (rakazo, OpenMausBot, OpenBot), agent teams (eigent, paperclip), single agents (openhuman, hermes-agent), and chat-native (nanoclaw, openclaw). Includes star counts and one-line positioning. Advice: pick the one that matches how you already work, run it on a real task for a week, keep only if it saves more time than it costs to host.
Never miss a night
One feed for everything, or one per agent harness you run. Rebuilt every night, with full credit on every item.
