Claude Memory Manager

A Claude Code skill that keeps your project's memory library searchable after months of accumulation.

English | 中文

---

Why

After a few months of work, the memory library is hard to search. You can't tell which entries are still current, which duplicate each other, or which file holds the answer to last quarter's bug.

Claude Code's auto-memory (v2.1.59+) writes plain markdown to ~/.claude/projects/<slug>/memory/ — you can read, edit, and version it. What it doesn't enforce is structure: naming, required fields, or a "why" section on each lesson. This skill adds those, plus a bash audit script that flags drift.

How It Works

• Schema on top of auto-memory. <type>_<topic>.md naming, required frontmatter (name / description / type), a Why section on feedback entries. Auto-memory still writes; the skill makes Claude write to a spec.
• Phrase-triggered review. "Audit memory" runs the script. "Review session" walks the recent session and surfaces what's worth keeping.
• Soft warning, no hooks. Audit reports drift; nothing blocks a write.
• Plain markdown on disk. Edit, grep, git-commit. The skill doesn't add a database or daemon.

Effect

• One topic per file means Claude lands on the right entry on the first lookup, not after several near-misses.
• A deduplicated library loads fewer files per session, freeing context for the work itself.

Sample audit output:

Memory audit · 2026-05-15 · 132 files

Hard checks (must be zero):
  missing frontmatter        0
  frontmatter fields         0
  feedback missing Why       1
  naming violations          0
  broken MEMORY.md links     0

Soft signals:
  oversized files           78
  groups over 15 entries     3
  untouched 30+ days        31
  not in MEMORY.md           0

Hard-rule compliance: 99.2%  (1 violation / 132 files)

Install

Tell Claude

Paste this into any Claude Code session:

Install the claude-memory-manager skill from
https://github.com/jau123/claude-memory-manager

Claude handles the rest. To verify, say "audit memory" in a new session.

Or install manually

git clone https://github.com/jau123/claude-memory-manager.git && \
  mkdir -p ~/.claude/skills/memory-management/templates && \
  cp claude-memory-manager/SKILL.md ~/.claude/skills/memory-management/ && \
  cp claude-memory-manager/templates/audit-memory.template.sh \
     ~/.claude/skills/memory-management/templates/

Per-project audit script + CLAUDE.md memory protocol → INSTALL.md

First Use

The skill activates from natural language. No slash command.

You: "Record today's wildcard bug fix"
→ Claude writes one feedback_*.md entry: filename, frontmatter,
  Why section, How-to-apply.

You: "Review the session"
→ Claude walks recent session, surfaces 3–5 candidates, asks
  which to keep.

You: "Audit memory"
→ Runs scripts/audit-memory.sh, reports compliance, lists files
  that need splitting.

Full trigger reference → SKILL.md

vs Built-in Auto-Memory

	Schema	Audit	Long-term result
Auto-memory alone	None (Claude decides)	None	Files accumulate without a naming or content spec
with this skill	3-type schema + required fields + Why on feedback	One-command script	Library stays auditable and searchable

For semantic retrieval over chunked storage, look at vector-backed tools like Mem0, Letta, or Zep.

Limits

• Single-project scope. One memory directory per skill instance; no cross-project consolidation.
• No semantic ranking. The audit is pattern matching (grep + filename + frontmatter); it won't catch "two files describe the same concept in different words."
• Bash + standard Unix tools. Tested on macOS bash 3.2 and Linux bash 5.x; Windows / git-bash untested.
• No concurrency safety. Don't run the audit while another session is mid-write.
• Overkill for small libraries. Below ~10 entries or a month of project age, the built-in auto-memory is sufficient and the schema overhead doesn't pay off.

License

MIT · Issues and PRs welcome at jau123/claude-memory-manager.