RecallLoom — Portable Project Memory

A file-based continuity harness that saves your project's current state, rules, and milestones so any session can pick up exactly where you left off.

AutomationIntermediate★ 154⑂ 5AI score 10/10Last updated: Aug 6, 2026

What it does

Maintains a .recallloom/ sidecar with three memory layers: context_brief.md (stable framing), rolling_summary.md (what is true right now), and daily_logs/ (append-only milestone evidence), plus state.json/config.json for machine-readable state.
Triggers on phrases like "continue this project", "restore project context", "pick up where we left off", and on stable action names rl-init, rl-resume, rl-status, rl-validate.
Enforces a cold-start read order (config → state → rolling summary → optional brief/daily log) and restores before doing work, so the agent judges instead of blindly continuing next_step.
All writes go through the dispatcher (write, append, record --plan/--suggest, sync-current-state-after-append) with layer classification, no_write/confirm/blocked exits, and hard bans on hand-editing state.json, config.json, or markers.
Ships four profiles (general, research writing, product docs, software coordination) and reference playbooks for recovery and file contracts.

Who it's for

Anyone running multi-day or multi-week projects with Claude Code who is tired of re-explaining context
Developers, researchers/writers, and PMs who need durable, human-readable project state
Teams switching between AI tools (Claude, Gemini, Copilot) that need one shared continuity layer

Examples

Run rl-init in a fresh repo to create the sidecar, validate the workspace, and get recommended next actions.
Return after a week and say "pick up where we left off" — resume --fast reads only state.json + rolling_summary.md to orient and propose the next safe move.
After shipping a release, say "record today's progress" — record --plan classifies the intent and append writes milestone evidence into today's daily log.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/Frappucc1no/recall-loom/HEAD/skills/recallloom/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)

Download the ZIP with the button below.
In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.

↓ Download ZIP

Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/recallloom folder from the GitHub repo Frappucc1no/recall-loom into my ~/.claude/skills/recallloom/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/Frappucc1no/recall-loom.git && mkdir -p ~/.claude/skills && cp -r recall-loom/skills/recallloom ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

Open a terminal and go to a folder where you keep source code.
Clone the repo: git clone https://github.com/Frappucc1no/recall-loom.git
Create the skills folder: mkdir -p ~/.claude/skills
Copy the skill in: cp -r recall-loom/skills/recallloom ~/.claude/skills/
Verify your runtime: python3 --version (3.10 or newer is required).
Restart Claude Code, open your project folder, and run rl-init (or ask "initialize project memory here").
In later sessions just say "continue this project" or rl-resume; use rl-status for a snapshot and rl-validate to check integrity.

View source on GitHub ↗License: Apache-2.0