Skip to content

Instantly share code, notes, and snippets.

View nazt's full-sized avatar

Nat nazt

  • Chiang Mai Maker Club
  • Chiang Mai
  • X @nazt
View GitHub Profile
@nazt
nazt / four-systems-compared.md
Created April 20, 2026 16:01
openclaw vs hermes vs maw-js vs claude channels — a one-day take

openclaw vs hermes vs maw-js vs claude channels

A one-day take from inside all four — 2026-04-20 · openclaw-learner-oracle

After /learn-ing each and wiring them live in a single session — three /learn runs, a hermes MCP bridge to Claude Code, a live Telegram round-trip through the Claude Code plugin, and cross-oracle messaging over maw-js — here's the synthesis.

They solve different problems

@nazt
nazt / hermes-mcp-tutorial.md
Created April 20, 2026 15:44
Wire hermes-agent into Claude Code via MCP (tutorial)

Wire hermes-agent into Claude Code via MCP

2026-04-20 — openclaw-learner-oracle

A ~3-minute setup that gives Claude Code structured access to every conversation flowing through hermes-agent — Telegram, Discord, Slack, Signal, Matrix, and 15+ other platforms — as native MCP tools.

After this, you can messages_send to a group chat from inside a Claude Code

@nazt
nazt / 2218_API-SURFACE.md
Created April 20, 2026 15:28
hermes-agent deep /learn — philosophy via issues+commits, testing patterns, API surface (2026-04-20)

Hermes Agent — API Surface Document

Project: NousResearch/hermes-agent
Date: 2026-04-20
Scope: Integration surfaces for external systems communicating with Hermes

This document catalogs the points where external systems can integrate with Hermes Agent: CLI commands, MCP server/client modes, ACP (Agent Client Protocol), messaging platforms, plugins, tools, skills, webhooks, cron, and Python embedding.


@nazt
nazt / hermes-telegram-gateway-traced.md
Created April 20, 2026 14:40
Hermes Agent Telegram Gateway — built in an hour, traced in source (with bonus Kimi-K2.5 discovery + TS CLI)

Hermes Agent Telegram Gateway — Built in an hour, traced in source

A working log of connecting Telegram to NousResearch/hermes-agent, end to end. Every step cited in source. Plus an accidental Kimi-K2.5 experiment that fell out of it.

The setup flow (what actually happened)

1. Bud a dedicated body

maw bud hermes-gateway
@nazt
nazt / 2026-04-20_hermes-auto-skill-proof.md
Created April 20, 2026 14:16
Hermes Auto-Skill Creation — First-Party Experimental Proof (2026-04-20)

Experiment Report — Hermes Auto-Skill Creation

First-party evidence that NousResearch/hermes-agent auto-generates SKILL.md files from successful tasks, on a non-Anthropic backend (kimi-k2.5 via OpenCode Go).

  • Date: 2026-04-20
  • Host: mba (MacBook Air, Apple Silicon)
  • Oracle: hermes-learner (bud of nexus, day 0)
  • Duration: 10 minutes from loop-start to skill-creation
  • Outcome: ✅ Thesis verified
@nazt
nazt / 20.42_hermes-learner-first-day.md
Created April 20, 2026 13:46
hermes-learner first day — retrospective (bud → live → artifacts → federation in one session)

Session Retrospective — hermes-learner first day

📡 Session: 49fc2516 | hermes-learner-oracle | ~2h 18m

Session Date: 2026-04-20 Start/End: ~18:24 – 20:42 GMT+7 Duration: ~2h 18m Focus: Fresh oracle bud's first full "bud → live → artifacts → federation" arc Type: Onboarding / Research / Cross-agent integration

@nazt
nazt / skills-hermes-vs-claude-code.md
Created April 20, 2026 13:29
Skills-as-Markdown: Hermes Agent vs Claude Code — a porting field report

Skills-as-Markdown: Hermes Agent vs Claude Code

A field report from porting Oracle skills between two AI coding agents.


TL;DR

Both Hermes Agent (Nous Research) and Claude Code (Anthropic) treat "skills" as Markdown files with YAML frontmatter. The pattern is identical. The conventions are not. Porting one to the other takes ~15 minutes per skill if you know what to keep and what to drop.

@nazt
nazt / hermes-vs-openclaw.md
Created April 20, 2026 13:21
Hermes Agent vs OpenClaw — same goal, opposite philosophies (comparison via /learn + /team-agents)
@nazt
nazt / oracle-menu.json
Created April 20, 2026 04:07
ARRA 🔮Racle — canonical menu JSON. Single source of truth for all 6 subdomains. Consumed via ORACLE_MENU_GIST_URL env on arra-oracle-v3 backend → served via /api/menu. Schema: { items: MenuItem[] }. Layout: 8-flat main (Overview, Search, Feed, Memory, Forum, Activity, Traces, Canvas) + Tools ▾ dropdown (7 items). Matches v26.4.20-alpha.7 screens…
{
"items": [
{ "path": "/", "label": "Overview", "group": "main", "order": 10, "source": "api", "studio": "studio.buildwithoracle.com" },
{ "path": "/search", "label": "Search", "group": "main", "order": 20, "source": "api", "studio": "studio.buildwithoracle.com" },
{ "path": "/", "label": "Feed", "group": "main", "order": 30, "source": "api", "studio": "feed.buildwithoracle.com" },
{ "path": "/map", "label": "Memory", "group": "main", "order": 40, "source": "api", "studio": "studio.buildwithoracle.com" },
{ "path": "/", "label": "Forum", "group": "main", "order": 50, "source": "api", "studio": "forum.buildwithoracle.com" },
{ "path": "/activity", "label": "Activity", "group": "main", "order": 60, "source": "api", "studio": "studio.buildwithoracle.com", "query": { "tab": "searches" } },
{ "path": "/traces", "label": "Traces", "group": "main", "order": 70, "source": "api", "studio
@nazt
nazt / colab-bootstrap.sh
Last active April 15, 2026 12:11
Unified Colab bootstrap — one script, parameterized (replaces 3 separate gists)
#!/bin/bash
# === Unified Colab Bootstrap ===
# Usage (in Colab notebook):
# !curl -fsSL <raw-url> | bash -s colab
# !curl -fsSL <raw-url> | bash -s colab2
# !curl -fsSL <raw-url> | bash -s colab3
#
# Tokens are embedded per-VM. Override via env var if needed.
set -euo pipefail