Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
Charles C. Figueiredo avatar

Agentcairn

  • 29 repo stars
  • Updated July 30, 2026
  • ccf/agentcairn

Local-first agent memory for Claude Code — recall, remember, ambient capture.

About

agentcairn is a Claude Code skill in the AI & Agent Building category. Local-first agent memory for Claude Code — recall, remember, ambient capture.

  • agentcairn
  • AI & Agent Building
  • AI-coding skill

Agentcairn by the numbers

  • Data as of Aug 1, 2026 (Skillselion catalog sync)
/plugin marketplace add ccf/agentcairn
/plugin install agentcairn@agentcairn

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
repo stars29
Last updatedJuly 30, 2026
Repositoryccf/agentcairn

What it does

Local-first agent memory for Claude Code — recall, remember, ambient capture.

README.md

agentcairn — Claude Code plugin

Local-first agent memory inside Claude Code: auto-wires the agentcairn MCP server, surfaces recent memory at the start of each session, and distills each session into a Markdown vault you own.

Install

claude plugin marketplace add ccf/agentcairn
claude plugin install agentcairn@agentcairn

On install you'll be asked for a vault path (default ~/agentcairn). The vault is auto-created (Obsidian-ready) on the first session — no Obsidian setup needed.

What you get

  • MCP tools: recall, search, build_context, recent, remember.
  • Ambient memory: SessionStart surfaces recent memories; SessionEnd distills the session.
  • Skill: using-agentcairn-memory (recall-before-work, remember durable facts).
  • Commands: /agentcairn:recall, /agentcairn:remember, /agentcairn:memory, /agentcairn:ingest.

The plugin runs the published agentcairn PyPI package via uvx — nothing to pip-install. You can also scaffold a vault yourself: uvx --from agentcairn cairn init ~/agentcairn.

Enabling the LLM judge

The plugin's background SessionEnd sweep reads ~/.agentcairn/config.toml directly — no shell-profile exports needed. To upgrade from the default embedding judge to the Anthropic LLM judge:

cairn config --init   # scaffold a fully-commented template (chmod 600)

Then uncomment two lines in ~/.agentcairn/config.toml:

judge = "anthropic"
anthropic_api_key = "sk-ant-..."

Related skills

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.