---
schema_version: "newruntime-agent-readable-v0.1"
type: "system_dossier"
slug: "hermes"
title: "Hermes"
description: "Hermes is tracked as a living experiment in memory, model routing, skills, delegation, local model lanes, and useful long-term personal agent behavior."
status: "published"
system_status: "running"
one_line_role: "Memory, model-routing, local-model, and self-improving agent laboratory."
research_question: "Can a personal agent become more useful through memory and skills without becoming opaque or unsafe?"
current_observation: "Hermes is the fast operational lane for interactive analysis and draft support; public claims should be tied to sanitized evidence."
verification_level: "workflow-tested"
last_verified: "2026-07-21"
related_shifts: ["shared-context-infrastructure","goal-scoped-agent-loops"]
related_posts: ["ai-native-development-stages","unabyss-shared-context-layer"]
topics: ["agent-systems","hermes"]
source_urls: []
public_boundary: "Sanitized dossier. Public routes intentionally exclude private operational commands, credentials, internal handoffs, and private discovery infrastructure."
routes: {"html":"https://newruntime.com/systems/hermes/","markdown":"https://newruntime.com/systems/hermes.md","json":"https://newruntime.com/systems/hermes.json"}
source_format: "markdown"
---

# Hermes

This public dossier is a placeholder for a sanitized system page. It should
track memory quality, routing, delegation, recoverability, and failure
containment without exposing private operational internals.

## Current public verdict

Hermes is a practical lab for fast newsroom and analysis operations. It becomes
public evidence only when a workflow is described without credentials, private
state, or sensitive deployment details.