Communityv1.0.0

Memory Scan

Scans OpenClaw agent memory files and workspace configs for malicious content, credential leaks, prompt injections, and security threats.

2kdownloads1stars5active installsdgriffin831
View on ClawHubBack to Skills

Skill Details

Slug
memory-scan
Latest Version
1.0.0
Author
dgriffin831
Published
Feb 2, 2026
Updated
May 11, 2026
Total Versions
1

How to Install

  1. 1 on OpenClawdBots (takes under 60 seconds).
  2. 2Open your bot dashboard and go to the Skills tab.
  3. 3Switch to the ClawHub tab and search for “Memory Scan”.
  4. 4Click Install and the skill is deployed to your bot automatically.

Changelog — v1.0.0

## [1.0.0] - 2026-02-01 ### Added - LLM-powered memory security scanner (`memory-scan.py`) - Detection of 8 threat categories: - Malicious Instructions, Prompt Injection, Credential Leakage, Data Exfiltration - Guardrail Bypass, Behavioral Manipulation, Privilege Escalation, Prompt Stealing - 5-level security scoring: SAFE (90-100), LOW (70-89), MEDIUM (50-69), HIGH (20-49), CRITICAL (0-19) - Quarantine system with backup and redaction (`quarantine.py`) - Scheduled scanning via cron job (`schedule-scan.sh`) - Multiple output formats: human-readable, JSON (`--json`), quiet (`--quiet`) - Single-file and full-workspace scanning modes - LLM provider auto-detection (OpenAI / Anthropic) - Eval framework with test cases (`evals/`) - Signal alerting integration for MEDIUM+ findings