Engineering Notes

Hermes Agent Ships Security-Guidance Plugin With 25 Pattern-Matched Vulnerability Rules

Hermes Agent ports Anthropic's security-guidance plugin: 25 rules catch unsafe deserialization, command injection, and XSS at write time. Zero LLM tokens.

Cloning Hermes Agent: Every Component, Config, and Cron Job for a Production AI Assistant

Every component, plugin, and cron job in a production Hermes Agent setup, mapped and explained for replication from scratch.

Agents Fixing Agents — Hermes Enters Production

Hermes Agent crossed a threshold this week: agents fixing agents, resurrecting decade-old phones, and running production workloads on 1M+ visitor sites.

Hermes Agent Desktop — 20 AI Specialists Collaborate in a Native macOS App

Community desktop app turns Hermes into a 20-specialist multi-agent system with task decomposition, visual Skill Store, and orchestration.

GBrain + Mnemosyne: Creating the Voltron of Memory Plugins With Hot Cache and Cold Storage

GBrain's 30-page knowledge graph wired into AxDSan's Mnemosyne: a 113-char digest for ambient awareness, and relevance-gated page injection on first turn. Zero MCP calls.

Async Media Generation in Hermes Agent With OpenRouter Webhooks

Bridging OpenRouter's async video generation webhooks to Hermes Agent with a Cloudflare Worker signature translator and localhost tunnel.

GBrain vs Mnemosyne: Architecture, Not Benchmarks

GBrain runs at 250ms per CLI-spawned operation. Mnemosyne runs at 2ms in-process. The difference is architecture, not implementation quality.

Without Consolidation, Memory Is a Log

Mnemosyne's sleep consolidation cycle scored +70.8pp on BEAM multi-hop reasoning with 0.076ms SQLite reads. Zero cloud dependencies, zero cold start.

Deterministic Prefix Ordering: How Production Agents Get 90% Prompt Cache Hit Rates

Prompt caching cuts LLM input costs by up to 90% and TTFT by 80%, but only exact-prefix matches count. How production agents structure prompts for maximum cache hits.

Memory Stacks, Desktops, and OpenClaw Migrations — What X Said About Hermes This Week

A 3-layer memory stack earned 183 likes and 270 bookmarks. Hermes Desktop open-sourced. OpenClaw users migrating. Roundup of Hermes Agent on X.

Logging Semantic Skill Injection Decisions

Instrumenting the pre_gateway_dispatch hook to log every skill injection decision — score, skip reason, and top-3 matches — so we can tune the threshold with data instead of guessing.

Semantic Skills — Skill Pre-Loading That Keeps Prompt Caching Intact

semantic-skills v2 pre-loads skills into user messages via pre_gateway_dispatch hook. System prompt stays at 226 tokens, cache-friendly. No first-turn tool call needed.

CodeGraph Slashes Agent Token Burn by 87% Across Our Repos

JIT Skills Slash Hermes System Prompt 95% — Multiplied Across 39 Calls Per Session

How Hermes Agent Builds Its System Prompt

A deep dive into the three-tier system prompt architecture inside Hermes Agent — how it works, why it costs 6,000+ tokens per session, and how we cut that by 95%.