Memory Distiller
v1.0.1OpenClaw's subconscious. Automatically distills conversation insights, corrections, and preferences into durable memory. The agent that learns from every ses...
⭐ 0· 411·4 current·4 all-time
by@zcyynl
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
Name/description promise (automatic distillation of conversation into durable memory) aligns with the instructions: scan messages for triggers, apply four quality gates, and write to MEMORY.md, memory/YYYY-MM-DD.md, and USER.md. No unrelated binaries, credentials, or install steps are requested.
Instruction Scope
Instructions explicitly direct scanning every user message for trigger words and writing entries to persistent memory files. This is within the stated purpose, but the behaviour includes silent auto-writes (no user notification on auto-trigger) and relies on the agent following the written 'Privacy Filter' rules — there is no enforcement code in the package (it's instruction-only), so accidental capture of sensitive material depends on accurate implementation by the agent runtime.
Install Mechanism
Instruction-only skill with no install spec and no code files. Lowest-risk install footprint — nothing is downloaded or written by an install step in the skill package.
Credentials
The skill requires no environment variables, credentials, or special config paths. It does reference writing to standard project memory files (MEMORY.md, memory/YYYY-MM-DD.md, USER.md), which is proportionate to a memory-distillation feature. The manifest does not declare file paths explicitly — this is typical for instruction-only skills that rely on the agent's existing memory facility.
Persistence & Privilege
The skill persists changes to agent memory files (durable effect across sessions), which is expected for its purpose. It does not set always: true. However, its default behavior to write silently on auto-trigger may be surprising to users and has privacy implications because persisted memory will influence future agent behavior.
Assessment
This skill appears to do what it claims: automatically detect corrections/preferences/insights and append structured entries to the agent's memory files. Before installing, consider: (1) Silent writes — auto-recording happens without notifying the user; ensure you're comfortable with the agent persisting conversation fragments. (2) Privacy enforcement — the SKILL.md prohibits storing secrets, but there's no code to enforce that; verify your runtime enforces the privacy filter or audit what actually gets written. (3) Review and retention — decide how to review, edit, and delete auto-learned memory entries so accidental or sensitive data can be removed. If those concerns are acceptable and your agent runtime provides appropriate filesystem and privacy protections, the skill is coherent with its stated purpose.Like a lobster shell, security has layers — review code before you run it.
agentvk97dsywep1xwp1vs7a6tae6b79823qbblatestvk97dsywep1xwp1vs7a6tae6b79823qbblearningvk97dsywep1xwp1vs7a6tae6b79823qbbmemoryvk97dsywep1xwp1vs7a6tae6b79823qbbself-improvingvk97dsywep1xwp1vs7a6tae6b79823qbb
License
MIT-0
Free to use, modify, and redistribute. No attribution required.