Persistent memory for Claude Code — identity, context, and continuity across sessions
# Add to your Claude Code skills
git clone https://github.com/Digital-Process-Tools/claude-rememberGuides for using ai agents skills like claude-remember.
Last scanned: 6/8/2026
{
"issues": [],
"status": "PASSED",
"scannedAt": "2026-06-08T08:55:45.318Z",
"npmAuditRan": true,
"pipAuditRan": true
}claude-remember is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by Digital-Process-Tools. Persistent memory for Claude Code — identity, context, and continuity across sessions. It has 165 GitHub stars.
Yes. claude-remember passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/Digital-Process-Tools/claude-remember" and add it to your Claude Code skills directory (see the Installation section above).
claude-remember is primarily written in Python. It is open-source under Digital-Process-Tools on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh claude-remember against similar tools.
No comments yet. Be the first to share your thoughts!
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.

Your coding agent starts every session blank. It doesn't know what you worked on yesterday, what conventions your team follows, or what mistakes it already made. You re-explain everything, every time.
Claude Remember fixes that. It hooks into your coding agent's lifecycle — saving sessions automatically, compressing them through Haiku into layered daily summaries, and loading them back into context on the next session start. No manual prompting, no copy-pasting notes. The agent starts every session with its history already present.
The result: your coding agent develops continuity. It remembers what it learned, what broke, what worked. Not perfect recall — compressed, practical memory that fits in minimal tokens.
flowchart TD
A["tool use"] --> B["save-session.sh"]
B --> C["extract (Python)"]
C --> D["summarize (Haiku)"]
D --> E["now.md"]
E --> F["hourly NDC compression"]
F --> G["today-YYYY-MM-DD.md"]
G --> H["daily consolidation"]
H --> I["recent.md + archive.md"]
Each layer compresses the one above it. Raw exchanges become one-line summaries. Daily summaries become weekly paragraphs. The result: full context in minimal tokens.
On session start, the SessionStart hook automatically injects into your coding agent's context:
identity.md — who the agent isremember.md — the handoff note from the last sessionnow.md — current session buffertoday-*.md — today's compressed historyrecent.md — last 7 daysarchive.md — older historyarchive-YYYY-MM-DD.md / recent-YYYY-MM-DD.md — rotated slices of a previously oversized archive or recent span; named at session start and searchable, but not injected into contextNo manual prompting, no "read this file" instructions. The agent begins every session with its memory already loaded. It just remembers.
After a compaction only identity.md is re-injected; the rest was already delivered. Write rules for the store: docs/how-memory-files-are-written.md.
Four plugins, one team, each does one thing. This one and three siblings:
All four install from one marketplace: /plugin marketplace add Digital-Process-Tools/claude-marketplace.
Once in a while, SessionStart names whichever of claude-supertool / claude-jit-context you have not installed yet, in a single systemMessage line the model never sees. It never speaks for a plugin already in ~/.claude/plugins/installed_plugins.json, and it stops entirely with "features": {"plugin_promos": false} in config.json: see Configuring it.
Claude Code
/plugin marketplace add Digital-Process-Tools/claude-marketplace
/plugin install remember@dpt-plugins
Restart Claude Code afterwards; hooks are read at session start (#200). Updating, the official Anthropic marketplace and its lag, manual install, checking your version: docs/install-claude-code.md.
Codex
codex plugin marketplace add Digital-Process-Tools/claude-remember
codex plugin install remember
Observed working against codex-cli 0.150.1. What was found on the way: docs/install-codex.md.
Antigravity CLI (agy)
python3 scripts/install_agy_hooks.py
Observed working against agy 1.1.27: capture was driven end to end against a real agy process, not reasoned from its docs. Antigravity has no per-plugin manifest -- agy plugin install copies a plugin's own hooks.json, counts it, marks the plugin enabled, and never loads it (#553) -- so the installer merges a remember entry into the shared ~/.gemini/config/hooks.json, preserving every other plugin's entries already there.
One gap, before you choose this host: of the four Antigravity events confirmed to fire, none is a process-exit signal, so there is no analogue of SessionEnd and no last-chance flush at the end of a conversation. Stop fires after every turn and is deliberately not wired to session-end-hook.sh. What that costs, the name-keyed schema whose parse failures are silent, and the three live defects found while porting: docs/install-antigravity.md.
claude) with Haiku accessjq and standard coreutils, preinstalled on macOS and LinuxNeeds a POSIX shell in PATH: Git Bash / MSYS2 with jq and python3 installed, or WSL. The OS badge is honest about the platform, not the coverage; most of the suite still skips on win32 (#497). Every real Windows defect so far was found by a user on a real machine, and those reports get priority. Known traps: docs/windows.md.
The pipeline uses Claude Haiku for summarization and compression. Haiku is the smallest, cheapest Claude model. A typical session save costs < $0.01 — a few thousand input tokens (the session exchanges) and a few hundred output tokens (the summary). Daily compression and consolidation add a few more Haiku calls.
In practice, running this all day costs a few cents per day. The Anthropic API key used by the Claude CLI is the same one that powers the calls — no separate billing.
Once installed there is nothing to run. Two commands are worth knowing.
/remember)Before clearing context or ending a session, type /remember. The agent writes a short handoff note; the next session starts with it loaded. How delivery is counted and what two sessions on one store do: docs/handoff.md.
/remember:doctor)Run it when memory is not appearing and nothing says why. It prints resolved paths, storage mode, and whether each hook has ever fired for this project. JSON output and the consolidation cap: docs/diagnostics.md.
| Claude Code / Codex | Antigravity | Script | Purpose |
|---|---|---|---|
SessionStart |
SessionStart |
session-start-hook.sh |
Loads memory into context, recovers missed sessions |
UserPromptSubmit |
PreInvocation (per model invocation, not per prompt) |
user-prompt-hook.sh |
Stamps the current time into the prompt |
PostToolUse |
not wired (PostInvocation fires, nothing here needs it) |
post-tool-hook.sh |
Saves the session when enough tool calls have accumulated |
SessionEnd |
no analogue found | session-end-hook.sh |
Flushes whatever PostToolUse has not saved yet |
Antigravity's Stop is a turn boundary, not a teardown, so it is wired to its own adapter rather than to session-end-hook.sh; the row above is the gap that leaves.
What each one skips and why, the hooks.d/ listener contract, and why SessionEnd never writes a handoff: docs/hooks.md.
Defaults live in config.json inside the plugin; override them per machine in ~/.remember/config.json and per project in <REMEMBER_DIR>/config.json. Every key, its default and what reads it: docs/configuration.md. Keeping memory outside the project tree, in ~/.remember/<slug>/: docs/external-storage-mode.md. Backing the store up to a git remote you own: docs/git-backup-security.md.
Everything lands in REMEMBER_DIR: .remember/ inside the project by default, or ~/.remember/<slug>/ in external storage mode.
| File | Purpose |
|---|---|
now.md |
Current session buffer |
today-*.md |
Daily compressed summaries |
recent.md |
Last 7 days consolidated |
archive.md |
Older history consolidated |
archive-YYYY-MM-DD.md, recent-YYYY-MM-DD.md |
Rotated slices, searchable, not auto-loaded |
remember.md |
Handoff note written by /remember |
identity.md |
Your agent's identity and values (you write this) |
logs/, tmp/ |
Local to this machine, never backed up |
Per-session handoff files, the session index, and the temp files tmp/ holds: docs/data-files.md.