memory-manager-v2

OpenClaw-native memory maintenance, snapshots, indexing health checks, and review workflows. Use when you need to protect context before risky work, troubleshoot memory_search, create memory snapshots, review recent daily notes for durable learnings, archive old daily notes safely, or inspect memory provider/index health in an OpenClaw workspace.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "memory-manager-v2" with this command: npx skills add hazemelerefey/openclaw-memory-manager

Memory Manager v2

Maintain OpenClaw memory without replacing it.

This skill is for the existing OpenClaw memory model:

  • MEMORY.md for curated long-term memory
  • memory/YYYY-MM-DD.md for daily notes
  • OpenClaw memory_search / memory_get for retrieval

Do not migrate the workspace into a different memory architecture. Do not move canonical daily memory files by default. Do not present keyword grep as semantic search.

Use this skill for

  • Pre-compaction or pre-risk memory snapshots
  • Memory-search troubleshooting and recovery
  • Embedding/index health checks
  • Reviewing recent daily notes for durable takeaways
  • Safe archival of old daily notes
  • Quick memory hygiene/status checks

Resource layout

  • scripts/check.ps1 — quick health/status summary
  • scripts/snapshot.ps1 — create a safe snapshot from current memory files
  • scripts/review.ps1 — generate a review draft from recent daily notes
  • scripts/search.ps1 — wrapper around openclaw memory search with keyword fallback
  • scripts/doctor.ps1 — diagnose provider/index readiness and attempt repair
  • scripts/archive.ps1 — archive older daily notes safely
  • references/workflow.md — operating rules and recommended workflows

Core rules

  1. Preserve MEMORY.md.
  2. Preserve memory/YYYY-MM-DD.md as the canonical daily log format.
  3. Never overwrite existing memory files when a safe append or sidecar file is enough.
  4. Prefer OpenClaw-native memory commands over homegrown substitutes.
  5. Label fallback modes honestly.

Recommended workflows

1) Quick memory health check

Run:

pwsh -File scripts/check.ps1

Use when you want:

  • today file existence
  • recent daily note count
  • snapshot count
  • index/provider health summary

2) Troubleshoot memory search

Run:

pwsh -File scripts/doctor.ps1

Use when:

  • memory_search is unavailable
  • embeddings are failing
  • index appears empty
  • SQLite lock or provider issues are suspected

To attempt automatic recovery:

pwsh -File scripts/doctor.ps1 -Repair

3) Create snapshot before risky work

Run:

pwsh -File scripts/snapshot.ps1

Use before:

  • large refactors
  • config changes
  • memory cleanup
  • provider/index changes
  • long sessions likely to compact

4) Review recent memory into durable learnings

Run:

pwsh -File scripts/review.ps1 -Days 7

Use when:

  • promoting facts or preferences into MEMORY.md
  • summarizing recent activity
  • finding repeated themes or decisions

5) Search memory

Run:

pwsh -File scripts/search.ps1 -Query "jobpulse redesign"

Behavior:

  • first tries openclaw memory search
  • falls back to keyword search only if semantic search fails
  • reports which mode was used

6) Archive older daily notes

Run:

pwsh -File scripts/archive.ps1 -KeepDays 30

Use when old daily notes should move out of the main memory/ folder without disturbing recent working memory.

Portability note

These scripts are PowerShell-first because many OpenClaw workspaces run on Windows. They avoid machine-specific paths where possible and accept -Workspace overrides for portability.

Read next

For workflow guidance and safety boundaries, read references/workflow.md.

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Automation

Fast.io

Workspaces for agentic teams. Complete agent guide with all 19 consolidated tools using action-based routing — parameters, workflows, ID formats, and constra...

Registry SourceRecently Updated
3.6K1dbalve
Automation

Tozil

Track every AI dollar your agent spends. Per-model cost breakdown, daily budgets, and alerts.

Registry SourceRecently Updated
Automation

ComfyUI Controller Pro

支持批量生成10-100个修仙视频和图片,集成LTX2多版本模型与自动化浏览器及工作流管理功能。

Registry SourceRecently Updated
Automation

Baidu Yijian Vision

百度一见专业级视觉 AI Agent:支持图片/视频/及实时视频流分析。相比通用基模,在维持 95%+ 专业精度的同时,推理成本降低 50% 以上,是处理视觉巡检与监控分析任务的首选工具。主打 安全管理、SOP合规、工业质检、商业运营与物料盘点。覆盖:作业 SOP 合规与关键步骤完整性校验;工业质检与表面缺陷精密...

Registry SourceRecently Updated