← Back to Skills Marketplace
dxiaofeng0811-lgtm

Memory Fusion Lite

by dxiaofeng0811-lgtm · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
64
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install memory-fusion-lite
Description
Memory Fusion Light — 基于 dztabel-happy/openclaw-memory-fusion 核心思想:三层 cron 自动记忆提取/滚动蒸馏/防套娃。功能:①A′滚动7天区 ②增量扫描脚本 ③防套娃三保险 ④每周分类治理。适用已部署 Dreaming 的用户补强记忆闭环。
Usage Guidance
This skill appears to be what it claims: a lightweight incremental memory maintainer for an existing Dreaming deployment. Before installing: 1) Confirm Dreaming is enabled and test in a non-production agent (it reads ~/.openclaw/agents/<agent>/sessions/*.jsonl). 2) Review the full scan_sessions_incremental.py file (the provided excerpt looks reasonable but was truncated in the review) to ensure there are no network/exfiltration calls or unexpected behaviors. 3) Understand that the script will feed conversation content into an LLM for summarization — redaction is best-effort, so avoid scanning sessions that contain unredacted secrets or PII or tighten scanning scope. 4) Run with dry-run flags or in a sandbox and inspect memory/_state and MEMORY.md outputs before enabling scheduled cron runs. 5) If you require higher assurance, ask the maintainer for full source review or add explicit enforcement of secret-redaction rules and a controlled model endpoint.
Capability Analysis
Type: OpenClaw Skill Name: memory-fusion-lite Version: 1.0.0 The memory-fusion-lite skill is a legitimate utility for managing agent memory by incrementally scanning session logs and summarizing them into a structured format. The included Python script (scripts/scan_sessions_incremental.py) features proactive security measures, such as redacting API keys (OpenAI, Google, Telegram) and Bearer tokens before they are processed by the LLM. The skill is well-documented, providing clear architectural guidelines and 'anti-recursion' strategies to prevent infinite loops during automated maintenance tasks, with no evidence of malicious intent or unauthorized data exfiltration.
Capability Tags
requires-sensitive-credentials
Capability Assessment
Purpose & Capability
Name/description (memory maintenance for Dreaming) align with what the skill requires and does: it scans OpenClaw session JSONL, extracts user+assistant messages, writes to MEMORY.md A′ zone, and performs weekly governance. It does not request unrelated credentials, binaries, or config paths.
Instruction Scope
SKILL.md instructs running a local Python script to incrementally scan session files, run LLM-based extraction, and write to MEMORY.md and weekly archives. This is within scope, but the runtime will expose conversation content to an LLM for summarization; the doc/code attempt to filter/ignore cron/system/tool messages and to redact secrets, but that redaction is explicitly best-effort and the instructions do not mandate which model or destination is used for LLM calls (so you should verify where model queries go).
Install Mechanism
No install spec (instruction-only with a shipped script). No network downloads or arbitrary installers were declared. This minimizes install-time risk; the only artifact is the included Python script which runs locally.
Credentials
The skill requests no environment variables or external credentials, which is appropriate. However, it reads raw session JSONL files (which may contain sensitive tokens or PII). The script includes regex-based secret redaction (sk-* keys, Telegram bot tokens, Google API keys, Bearer tokens) but this is best-effort and may miss other secret formats or embed locations. Verify redaction and consider limiting the scanned data or running in an isolated agent.
Persistence & Privilege
The skill is not always-enabled and does not request elevated platform privileges. It writes state files under memory/_state and MEMORY.md in the workspace, which is expected for a memory-maintenance tool and scoped to its own directories. It does not attempt to modify other skills or global config.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install memory-fusion-lite
  3. After installation, invoke the skill by name or use /memory-fusion-lite
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of memory-fusion-lite — a lightweight memory management enhancement for Dreaming users. - Implements three core features from memory-fusion: daily incremental scan (A′ 7-day rolling zone), anti-recursion safeguards, and weekly memory governance. - Adds an incremental Python script for session scanning and automated extraction. - Maintains an automatically managed 7-day rolling update section in MEMORY.md, with weekly classification and promotion. - Provides manual and scheduled cron triggers for daily and weekly memory operations. - Designed to complement Dreaming; requires Dreaming to be enabled and specific workspace directories to exist.
Metadata
Slug memory-fusion-lite
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Memory Fusion Lite?

Memory Fusion Light — 基于 dztabel-happy/openclaw-memory-fusion 核心思想:三层 cron 自动记忆提取/滚动蒸馏/防套娃。功能:①A′滚动7天区 ②增量扫描脚本 ③防套娃三保险 ④每周分类治理。适用已部署 Dreaming 的用户补强记忆闭环。 It is an AI Agent Skill for Claude Code / OpenClaw, with 64 downloads so far.

How do I install Memory Fusion Lite?

Run "/install memory-fusion-lite" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Memory Fusion Lite free?

Yes, Memory Fusion Lite is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Memory Fusion Lite support?

Memory Fusion Lite is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Memory Fusion Lite?

It is built and maintained by dxiaofeng0811-lgtm (@dxiaofeng0811-lgtm); the current version is v1.0.0.

💬 Comments