Triple Memory Lake

Quad-layer memory system integration - unifies OpenClaw, Claude Code, and self-improving agent memories into a single knowledge lake

This skill will copy files from your home directory (~/.claude and ~/.openclaw) into a local memory/ folder and then process them to extract patterns. Before installing or enabling autonomous execution: (1) Inspect the memory/ and scripts locally to see exactly what will be copied and generated. (2) Consider running the sync scripts manually in an isolated environment (or container) first to review outputs and ensure no secrets are being copied. (3) If you do want this behavior, restrict execution to manual invocation or audit the files to redact secrets; if you expect OpenClaw daily logs to be synced, note that the skill currently lacks a script for that and may not be doing what you expect. (4) If you are uncomfortable with duplicating potentially sensitive logs, do not enable autonomous invocation and consider removing or modifying the scripts to filter/redact sensitive fields before saving.

Type: OpenClaw Skill Name: triple-memory-lake Version: 1.1.0 The skill is designed to aggregate sensitive session logs and metrics from other AI tools by reading directly from `~/.claude` and `~/.openclaw` (found in sync-claude-code.py and sync-self-improving.py). While this behavior aligns with the stated goal of creating a 'unified memory lake,' the broad access to private interaction history and the use of subprocesses to orchestrate these syncs represent a high-risk capability. No evidence of remote exfiltration was found, but the automated collection of local session data warrants caution.