Learnloop

PassAudited by ClawScan on May 14, 2026.

Overview

LearnLoop is a coherent local memory skill, but installing it means selected corrections, preferences, and project facts may be stored and reused in future sessions.

Install this only if you want Claude to keep local memories across sessions. Do not let it save secrets, credentials, or sensitive personal/business details unless you intentionally want them retained, and periodically inspect or clean `~/.claude/projects/<id>/memory/` and `MEMORY.md`.

Findings (1)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Low

#ASI06: Memory and Context Poisoning

What this means

Corrections, preferences, and project facts may remain in local Claude memory and shape future sessions without needing to be re-entered.

Why it was flagged

The skill intentionally stores memories that will be loaded into future sessions, so incorrect, outdated, or sensitive entries could persistently affect later agent behavior.

Skill content

Persist — writes to Claude Code's native auto-memory at `~/.claude/projects/<id>/memory/`, auto-injected into every future session

Recommendation

Review the memory directory periodically, avoid saving secrets or highly sensitive information, and ask the agent to confirm before saving anything private or long-lived.