honcho

Configure Honcho cross-session memory with peer isolation and dialectic reasoning for Hermes.

14|5|Updated Apr 9, 2026
One-click install
npx skills add https://github.com/MLT-OSS/hermes-agent-go --skill honcho-mlt-oss
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: honcho
Source: https://github.com/MLT-OSS/hermes-agent-go/tree/main/optional-skills/autonomous-ai-agents/honcho
Command: npx skills add https://github.com/MLT-OSS/hermes-agent-go --skill honcho-mlt-oss

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires honcho-ai.

What problem does it solve? AI agents forget users between sessions, forcing people to repeat preferences and context. This Skill sets up Honcho memory for Hermes so user profiles, observations, and conclusions persist across conversations and profiles. ## Core Features & Use Cases - Cross-Session User Modeling: Honcho builds a persistent representation of each user from observed messages, recalled via profile, search, context, and conclude tools. - Multi-Profile Peer Isolation: Each Hermes profile gets its own AI peer while sharing one workspace, so agents develop independent views of the same user. - Tunable Observation & Recall: Configure observeMe/observeOthers toggles, recall modes (hybrid, context, tools), session strategies, and dialectic reasoning levels. - Use Case: Create a coding profile with hermes profile create coder --clone so it gets its own Honcho peer, then tune its recall mode to tools for explicit memory control. ## Quick Start Run the Honcho setup wizard with hermes honcho setup, choose cloud or self-hosted, and verify the connection with hermes honcho status.

Frequently Asked Questions about honcho

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up Honcho memory with Hermes?

Run hermes honcho setup and choose cloud (paste an API key from app.honcho.dev) or local (enter a self-hosted base URL). Verify the connection and resolved config with hermes honcho status.

How do I give each Hermes profile its own Honcho peer?

Create profiles with hermes profile create <name> --clone, which adds a host block and sets aiPeer to the profile name. For existing profiles, run hermes honcho sync to backfill host blocks.

What is the difference between hybrid, context, and tools recall modes?

Hybrid auto-injects context and exposes the four Honcho tools. Context mode only auto-injects memory with no tool calls, minimizing token cost. Tools mode hides auto-injection so the agent controls all memory access explicitly.

Why is Honcho memory not persisting across sessions?

Check hermes honcho status and confirm saveMessages is true and writeFrequency is not set to session, which only writes on exit. Also verify memory.provider is set to honcho in the Hermes config.

Can I use Honcho with a self-hosted server instead of the cloud?

Yes, select local during hermes honcho setup and provide your base URL, such as http://localhost:8000. The same peer, observation, and recall configuration applies to self-hosted deployments.

Why are my Honcho observation settings from the dashboard not applied?

Observation config syncs from the server on each session init, and server-side settings override local defaults. Start a new Hermes session after changing toggles in the Honcho dashboard.