Security audit

Moltbook Weekly Report

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed weekly-report generator that reads limited local project data and prepares WeCom publishing guidance, with review-before-sharing caveats.

Before installing, treat generated reports as potentially containing internal operational data. Review the Markdown before running any WeCom command, confirm the destination document/workspace and visibility, and do not assume --auto-publish actually published anything unless you verify it in WeCom.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The script reads local operational data from /tmp/v35_migration_config.json and /tmp/price_monitor_db.json and embeds it into a generated report without any validation, consent boundary, or clear declaration in the interface. In a skill that is ostensibly for weekly report generation and WeCom publishing, this creates an information disclosure risk because unrelated local state can be silently exfiltrated into output that may be shared externally.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The README advertises automatic collection of project data and one-click publishing to an enterprise WeCom document, but it does not warn users that potentially sensitive operational or business data may be transmitted to an external service. In a reporting skill, this omission can lead to unintentional disclosure because users may enable auto-publish without understanding what data sources are included or where the content is sent.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The quick-start instructions promote automatic publishing to Enterprise WeChat without warning that generated content will be transmitted to an external service and may create persistent documents. Users may run the command expecting local report generation and unintentionally disclose internal operational data to a broader audience.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill describes automatic aggregation of several operational data sources without caution about sensitivity, access control, or human review before publication. Combining metrics from production systems and analytics sources increases the risk of exposing confidential business data, performance details, or internal operational patterns in generated reports.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The one-click publishing section omits that report contents will be written into Enterprise WeChat and that a share link may be generated, expanding access beyond the local environment. This is dangerous because a single action can create durable, externally accessible copies of sensitive internal reporting data with little user awareness.

Static analysis

No suspicious patterns detected.