Openclaw Memory Enhancer
PassAudited by ClawScan on May 1, 2026.
Overview
The artifacts describe a coherent local memory/RAG helper, but users should understand that it persists recalled context across sessions and includes optional external setup paths.
This appears purpose-aligned for local long-term memory. Before installing, decide what should be allowed into the memory folder, review remembered content periodically, and be cautious if your OpenClaw setup sends prompts to a remote LLM. If you use the Git or standard-version setup, inspect the upstream code and dependencies first.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Information saved in memory may be reused in future conversations and could steer the agent's answers.
The skill is designed to persist, retrieve, and inject memory content into future prompts. This is central to the memory-enhancement purpose, but sensitive or poisoned memories could influence later responses.
Automatically reads all files from `memory/` directory ... `instruction` | Direct instructions | "Remember: always do X" ... `enhanced_prompt` includes `{memory_context}` before `User: {user_input}`Only store intended information in the memory directory, periodically review/delete memory entries, and ensure recalled memories are treated as context rather than authoritative instructions.
If you follow the Git or standard-version setup, you may run code or dependencies that were not included in the supplied artifact review.
The documentation gives user-directed paths to fetch upstream code and install unpinned Python packages. This is purpose-aligned setup guidance, but it is outside the instruction-only artifact itself.
git clone https://github.com/henryfcb/openclaw-memory-enhancer.git ... ~/.openclaw/skills/openclaw-memory-enhancer ... pip install sentence-transformers numpy
Inspect the upstream repository before running it, prefer pinned versions or reviewed commits, and use the edge version if you want to avoid extra dependencies.