Ultimate AI agent memory system for Cursor, Claude, ChatGPT & Copilot. WAL protocol + vector search + git-notes + cloud backup. Never lose context again. Vibe-coding ready.
Security Analysis
high confidenceThe skill's requirements and instructions are consistent with a scaffolding/documentation-first memory system: it creates local memory files, recommends vector/third-party services, and asks only for an OpenAI key; nothing in the package appears disproportionate or covert.
Name/description promise (WAL + vector search + git-notes + optional cloud backup) matches what the skill actually provides: templates, instructions, and a small CLI to initialize local files and recommend integrations. The single required env var (OPENAI_API_KEY) aligns with the documented use of OpenAI as the memory search provider.
SKILL.md instructs the agent/user to create SESSION-STATE.md, edit agent config files (~/.openclaw/openclaw.json, ~/.clawdbot/clawdbot.json) and optionally configure third-party services (LanceDB, Mem0, SuperMemory). Those actions are within scope for a memory system but the instructions reference external tools (python memory.py, memory_recall, mem0, supermemory) that are not bundled — user will need to install/configure them separately. No instructions attempt to read unrelated system secrets or exfiltrate data.
There is no install spec (instruction-only style) and the package contains a small benign CLI (bin/elite-memory.js) that writes/reads local markdown files and checks a local path for LanceDB. No remote downloads or extracts are performed by the skill itself. package.json lists mem0ai as an optional dependency, which is reasonable for optional auto-extraction functionality.
Only OPENAI_API_KEY is declared as required, which matches the documented use when enabling OpenAI-backed memory search. The SKILL.md mentions additional optional API keys (MEM0_API_KEY, SUPERMEMORY_API_KEY) for third-party features; these are optional and not required by the CLI. This is proportionate, but users should be aware that enabling optional features will require providing those extra credentials.
The skill does not request always:true and does not attempt to modify other skills or system-wide agent settings programmatically. Its CLI writes files into the current workspace and reads HOME to inspect a local LanceDB path — expected behavior for this kind of tool.
Guidance
This package is primarily a scaffolding and documentation bundle with a small CLI that creates local memory files (SESSION-STATE.md, MEMORY.md, daily logs). Before installing: 1) Review the GitHub/homepage and publisher identity (registry metadata lists owner but source was noted as unknown). 2) Back up any agent config files (~/.openclaw/openclaw.json, ~/.clawdbot/clawdbot.json) before editing them. 3) Understand that optional features (Mem0, SuperMemory, LanceDB) require additional API keys and external services you must trust. 4) Inspect bin/elite-memory.js (it only writes/reads local markdown files and checks a local path) and the SKILL.md to confirm you want the agent to create persistent local memory files. If you plan to enable cloud backups or third-party memory providers, verify those providers' privacy/security practices before supplying keys.
Latest Release
v1.2.3
- Expanded description and keywords for improved discoverability, highlighting ChatGPT, Copilot, Cursor, and developer tooling. - Updated version to 1.2.3. - Clarified compatibility and marketing text in SKILL.md. - No functional or implementation changes to code logic.
More by @NextFrontierBuilds
Published by @NextFrontierBuilds on ClawHub
