Hermes Labyrinth - Setup Guide¶

Source: reason-machines/trending-skills (187+ installs) Category: Agent Infrastructure / Observability Quality Tier: 🔵 Community

A read-only observability dashboard plugin for Hermes Agent. Turns autonomous agent runs into a navigable map of crossings (prompts, tool calls, tool results, failures, model switches, subagents, approvals, memory hits, redactions, context compression, cron runs) with exportable evidence.

Installation¶

Plugin Directory Install¶

mkdir -p ~/.hermes/plugins
git clone https://github.com/stainlu/hermes-labyrinth.git ~/.hermes/plugins/hermes-labyrinth

Start or restart the Hermes dashboard:

hermes dashboard

If the dashboard is already running, rescan plugins:

curl http://127.0.0.1:9119/api/dashboard/plugins/rescan

Open the dashboard in your browser and select the Labyrinth tab.

Optional Theme¶

mkdir -p ~/.hermes/dashboard-themes
cp ~/.hermes/plugins/hermes-labyrinth/theme/hermes-labyrinth.yaml ~/.hermes/dashboard-themes/

What Labyrinth Tracks¶

View	Contents
Journey index	Recent CLI, dashboard, gateway, cron, and delegated work
Labyrinth map	Ordered crossings through a selected agent journey
Inspector	Input, output, duration, status, evidence, guideposts per crossing
Guideposts	Generated observations backed by local evidence
Skill atlas	Bundled, optional, external, and user skill inventory
Cron gate	Scheduled autonomy, next runs, last failures, workdirs
Model ferry	Model/provider transitions across sessions
Reports	Redacted Markdown and JSON exports for one journey

API Reference¶

All endpoints are read-only. The plugin API is served at:

http://127.0.0.1:9119/api/plugins/hermes-labyrinth/

Endpoints¶

GET /health
GET /journeys
GET /journeys/{journey_id}
GET /journeys/{journey_id}/crossings
GET /skills
GET /cron
GET /guideposts
GET /reports/{journey_id}.json
GET /reports/{journey_id}.md

Example: Fetch All Journeys¶

curl http://127.0.0.1:9119/api/plugins/hermes-labyrinth/journeys | jq .

Example: Export a Journey Report¶

JOURNEY_ID="your-journey-id"
curl "http://127.0.0.1:9119/api/plugins/hermes-labyrinth/reports/${JOURNEY_ID}.md" > report.md

Health Check¶

curl http://127.0.0.1:9119/api/plugins/hermes-labyrinth/health

Python API Client¶

import urllib.request
import json

BASE = "http://127.0.0.1:9119/api/plugins/hermes-labyrinth"

def get_journeys():
    with urllib.request.urlopen(f"{BASE}/journeys") as r:
        return json.loads(r.read())

def get_crossings(journey_id: str):
    with urllib.request.urlopen(f"{BASE}/journeys/{journey_id}/crossings") as r:
        return json.loads(r.read())

def get_report_json(journey_id: str):
    with urllib.request.urlopen(f"{BASE}/reports/{journey_id}.json") as r:
        return json.loads(r.read())

Verification¶

# After installing plugin and starting dashboard
curl http://127.0.0.1:9119/api/plugins/hermes-labyrinth/health

# Should return: {"status": "ok"}

# Check journeys endpoint
curl http://127.0.0.1:9119/api/plugins/hermes-labyrinth/journeys | jq '. | length'

Notes¶

Read-only by design - no write operations, safe to run in production
Black-box recorder - not a chat UI, designed for post-hoc analysis
Exportable evidence - JSON and Markdown reports for sharing findings
From the ara.so Daily 2026 Skills collection
Complements openclaw-control-center for agent monitoring