Claude Remember

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.

How it works

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 is
• remember.md — the handoff note from the last session
• now.md — current session buffer
• today-*.md — today's compressed history
• recent.md — last 7 days
• archive.md — older history
• archive-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 context

No 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.

From the same workshop

Four plugins, one team, each does one thing. This one and three siblings:

• claude-jit-context: project knowledge that loads only when the prompt, the file or the tool matches it.
• claude-supertool: batched file and tracker ops. One call instead of seven, and a refusal instead of a wrong answer.
• claude-oss: the maintainer loop that runs these four repos. Triage, build, review, merge, release.

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. The line opens with claude-remember: and ends with that key, so both who spoke and how to stop it arrive with the message rather than only here (#631).

Install

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