Security audit

Daily Memory Summary

Security checks for vulnerabilities and agentic risk

Overview

The skill has an understandable purpose, but it automatically stores sensitive clipboard and notification content in persistent memory with weak controls and some under-disclosed contact parsing.

Install only if you intentionally want daily clipboard and notification records saved into persistent OpenClaw memory. Before enabling it, review or remove any scheduled cron entry, change the hard-coded workspace path, consider deleting the contact-parsing block, and add redaction or review steps for secrets, private messages, one-time codes, and work data.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 89% confidence
Finding: The documented purpose says the skill summarizes clipboard and notification data into a daily memory, but the finding indicates it also parses identity information from notifications and updates a contacts knowledge base in a separate directory. That hidden expansion of scope is dangerous because it enables collection and persistence of personal relationship data beyond the user's stated expectation, reducing transparency and informed consent.

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The skill's stated purpose is to summarize clipboard and notification data into daily memory, but the implementation also mines notification content for contact identities and claims to update a contacts library. This scope expansion increases collection and processing of personal data beyond the declared behavior, creating a transparency and least-privilege violation that can lead to unauthorized profiling or retention.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: This skill processes highly sensitive sources—clipboard contents and notifications—and writes the resulting data into a persistent daily memory file, yet the description does not clearly warn the user. Missing disclosure is dangerous because users may unknowingly authorize capture and retention of secrets, private messages, one-time codes, or other sensitive content from transient system surfaces.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The script reads clipboard contents, which commonly contain passwords, tokens, personal messages, or confidential text, and does so without any explicit consent prompt, warning, or sensitivity check. In this skill context, the data source is highly sensitive computer I/O, so silent ingestion materially increases privacy and data-handling risk.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The script persists raw clipboard and notification content to a markdown file on disk without warning the user that transient data will become a durable record. This increases exposure because any local compromise, backup sync, later indexing, or accidental sharing can reveal data that the user did not expect to be retained.

Ssd 3

Medium

Confidence: 96% confidence
Finding: The script copies raw clipboard and notification contents directly into a persistent daily memory file, creating a durable data retention channel for sensitive user-provided and incoming information. Because the skill operates on computer I/O artifacts that may include credentials, private conversations, corporate information, and personal identifiers, persistent storage materially raises confidentiality risk.

Ssd 3

Medium

Confidence: 87% confidence
Finding: The described workflow encourages broad collection from clipboard and notification sources into memory and contact-related processing without clear minimization boundaries. In this context, that design increases the chance of over-collection of sensitive data unrelated to the task, which is a privacy and governance weakness even if no immediate exploit primitive is present.

Static analysis

No suspicious patterns detected.