honcho-memory — Claude Code skill: memory and setup
The honcho-memory skill explains how a connected Honcho instance stores user context and feeds it back to an agent. It fits conversations that need continuity across sessions.
AGPL-3.0 214 installs ★ 8,222 updated 2026-10-23
npx skills add plastic-labs/honcho --skill honcho-memory -a claude-codeWhat honcho-memory does
honcho-memory lays out a working loop for Honcho memory: retrieve relevant context before a reply, produce the response, then save both sides of the exchange. Honcho reasons over recorded messages and builds a user representation that carries across sessions.
The skill covers peers, sessions, messages, conclusions, and representations. Use get_representation, get_session_context, or search for quick reads; use chat when a question needs a reasoned answer. Its reasoning level runs from minimal to max, while representation updates happen asynchronously.
Honcho must already be connected, and an API key beginning with hch- is required. For Claude Code, install the skill with the commands below; a raw MCP connection can use https://mcp.honcho.dev.
What it helps with
- retain user context across conversations
- retrieve preferences before generating a reply
- record both sides of each exchange
- keep separate conversation and project sessions
- ask Honcho for a reasoned user profile
Example prompts for the agent
- Read the user’s current context with
get_session_contextbefore answering. - Use the user representation to tailor this reply to prior preferences.
- After replying, record the user message and your response in Honcho.
How to install honcho-memory
1Skills CLI (Vercel)
npx skills add plastic-labs/honcho --skill honcho-memory -a claude-code2Manually: put the files in place
~/.claude/skills/honcho-memory/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/honcho-memory/
1Skills CLI (Vercel)
npx skills add plastic-labs/honcho --skill honcho-memory -a codex2Manually: put the files in place
.agents/skills/honcho-memory/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add plastic-labs/honcho --skill honcho-memory -a cursor2Manually: put the files in place
.cursor/skills/honcho-memory/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install plastic-labs/honcho skills/honcho-memory2Manually: put the files in place
.github/skills/honcho-memory/SKILL.md1Skills CLI (Vercel)
npx skills add plastic-labs/honcho --skill honcho-memory -a gemini-cli2Platform CLI
gemini skills install https://github.com/plastic-labs/honcho --path skills/honcho-memoryManually: put the files in place
.devin/skills/honcho-memory/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add plastic-labs/honcho --skill honcho-memory -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about honcho-memory
How do I install honcho-memory in Claude Code?
Run npx skills add plastic-labs/honcho --skill honcho-memory -a claude-code, then check ~/.claude/skills/honcho-memory/SKILL.md.
What does honcho-memory do?
It defines the Honcho recall, response, and recording cycle. The agent retrieves context, answers, and stores the exchange so the user representation can grow across sessions.
What does honcho-memory require?
A connected Honcho instance and an API key with the hch- prefix are required. Raw MCP access can point to https://mcp.honcho.dev.