OpenClaw Memory System
v1.0.0A lightweight, prompt-driven three-layer memory system for long-term context retention in OpenClaw agents with secure, asynchronous data extraction and Markd...
⭐ 0· 47·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The skill claims to provide a three-layer L1/L2/L3 memory system and all artifacts (SKILL.md, prompts, templates, references) implement that design. Required paths, cron registrations, backup/restore, and file-based persistence are coherent with the stated purpose. No unrelated credentials, network endpoints, or binaries are requested.
Instruction Scope
Runtime prompts instruct the agent to read/write/append files under a workspace and to scan the agent sessions directory (../../agents/main/sessions/). They also define a GC deletion policy (delete files containing ':run:' or <1KB). These operations are consistent with memory management, but they are destructive when misconfigured: if BASE/pwd is different than expected, or the sessions path layout differs, deletion/overwrites could affect unintended files. Prompts reference agent 'read'/'write' tools (assumed platform tools) but do not explain their permissions — review how those tools behave in your environment.
Install Mechanism
No install spec and no external downloads — instruction-only plus one local validation shell script. This is low-risk from an installation standpoint; nothing is fetched from arbitrary URLs or written outside the skill/workspace except via Cron registration (which is expected).
Credentials
The skill requests no environment variables, no credentials, and no config paths outside the agent workspace. All file and path accesses declared by prompts relate to the memory system (workspace, memory/, SESSION-STATE.md, sessions/). There are no demanded API keys or unexplained secret access.
Persistence & Privilege
The skill registers Cron tasks (openclaw cron add) to run prompts on a schedule and will therefore create persistent, periodic behavior. This is appropriate for a memory system but does increase runtime persistence. always:false and no modifications to other skills' configs are observed. Confirm you are comfortable with scheduled autonomous tasks that will read/write session files.
Assessment
This skill appears to be what it says: a file-based, prompt-driven memory system that schedules Cron tasks and manipulates local workspace and sessions files. Before installing: (1) verify the BASE/pwd layout used in prompts matches your OpenClaw installation (the prompts assume a specific ../../agents/main/sessions path); (2) review and, if needed, adapt the GC deletion rules (':run:' and <1KB criteria) to avoid accidental deletion in your environment; (3) confirm what the platform 'read'/'write' tools do and what permissions they hold; (4) ensure you are comfortable with automatic Cron registrations (they will persist and run autonomously); (5) test in a non-production agent or backup your workspace and run the included validate.sh locally to see expected file effects. If any of these items are unclear or you need the skill to be less aggressive, request changes prior to installation.Like a lobster shell, security has layers — review code before you run it.
latestvk97bwyhapt2f46wr4r6s51t1pd84r19h
License
MIT-0
Free to use, modify, and redistribute. No attribution required.