Memory Fortress

Complete memory management system for OpenClaw agents. Combines compaction-aware saving, a formal boot sequence, domain organization, memory scoring, structu...

This skill appears to do what it says: it organizes and persists agent memory to structured files. Before installing, consider privacy and retention: the agent is instructed to save users' exact words and other session content to disk, so sensitive information (passwords, API keys, PII) could be permanently stored unless you put controls in place. Recommended precautions: (1) confirm agent workspaces are isolated and not auto-synced to external cloud storage or shared repositories; (2) add or enable automatic redaction/filters so the agent strips PII/secrets before writing; (3) set retention/archival policies (auto-delete or encryption-at-rest) for memory files; (4) review multi-agent behavior to ensure it can't read other agents' workspaces; and (5) test in a sandboxed environment first. If you need the skill but want stronger guarantees, ask the author to include explicit redaction rules and retention controls in SKILL.md.

Type: OpenClaw Skill Name: memory-fortress Version: 1.1.2 The 'memory-fortress' skill is a comprehensive organizational framework for AI agents to manage long-term memory and project state using local Markdown files. It defines a structured directory system (state/, memory/, projects/) and provides clear instructions for the agent to document tasks, decisions, and learnings to ensure continuity across sessions. The installation steps are limited to benign filesystem operations (mkdir, touch), and there is no evidence of data exfiltration, malicious execution, or harmful prompt injection.