Hierarchical Agent Memory

• Risk level: Low
• Analyzed: Apr 11, 2026, 6:20 PM
• Findings: 0
• Version: 3.1.0

Overview

The skill's instructions, required artifacts, and behavior are coherent with its stated purpose of organizing agent memory; it asks for no external credentials, performs only workspace reads/writes, and has no install footprint.

This skill is internally consistent with what it says it will do, but review these points before installing: 1) Back up your existing MEMORY.md and memory/ directories (or ensure git) so you can revert migrations. 2) The agent will read openclaw.json and workspace files to detect upgrades/plugins — ensure you are comfortable with that. 3) Never store secrets or API keys in MEMORY.md or any files pointed to from it; follow the skill's recommendation to keep credentials in environment variables or .env (excluded from memory). 4) During onboarding, the agent will ask for explicit approval before creating or modifying files — decline or review changes if you want manual control. 5) If you enable the 'full' preset, verify the cron/distillation schedules and timezone settings. If you want extra assurance, inspect the onboarding conversation the agent runs and the exact migration plan it proposes before letting it execute any file writes.