Security audit

memory-system

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed persistent-memory framework that writes local memory files, with privacy cautions and no evidence of exfiltration or hidden behavior.

Install this only if you want the agent to maintain persistent local memory. Review what gets written to MEMORY.md and memory/ before relying on it, avoid storing secrets or regulated personal data unless you explicitly intend that, and do not enable cron automation unless you are comfortable with unattended local memory updates.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (4)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The guidance tells the agent to persist decisions, context, and other remembered information to disk, but it does not require explicit user consent or a clear warning that stored data may include sensitive personal information and survive future sessions. Even though it says to skip secrets unless asked, that is too weak as a safeguard because agents may still persist private or regulated data by mistake.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The maintenance workflow instructs the agent to routinely read, distill, update, remove, and archive memory files, which gives it authority to reorganize or alter user data without an explicit warning, approval boundary, or audit expectation. In practice this can lead to silent data retention, accidental deletion, over-collection, or movement of sensitive content into long-term stores that users did not intend.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The document instructs the agent to modify and archive persistent memory files (including rewriting MEMORY.md and moving sections to archive files) without explicitly requiring user awareness or confirmation. In a memory-management skill, silent persistent changes can alter future agent behavior, destroy context visibility, or preserve sensitive information in unexpected locations, making this a real safety issue even if the feature is intentional.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The optional cron automation section describes scheduled execution that will periodically read, update, archive, and log memory state without a clear warning about unattended persistent file changes. Because this skill's purpose is long-lived memory management, unattended automation increases the risk of unnoticed data retention, corruption, privacy leakage, or irreversible memory drift over time.

Static analysis

No suspicious patterns detected.