Back to skill

Security audit

WeChat Sender Windows

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says, but it can automatically send WeChat messages from a logged-in account with weak recipient and window verification.

Review carefully before installing. Use only for contacts and messages you are willing to verify manually, avoid batch sending, and do not run it with sensitive message text unless the recipient and WeChat window are visibly confirmed first. Prefer a version with exact recipient matching, preview/confirmation, escaped literal text input, and pinned dependencies.

Vulnerability Patterns
  • Tool Hijacking and SpoofingModifies or replaces tools so legitimate-looking calls execute attacker logic
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Findings (4)

T09 · Insecure Skill Coding Practices

Error
Location
send_wechat.py:91
Finding

User-Controlled Keystroke Sequence Injection

Content
View full analysis
Remediation
View remediation

T07 · Tool Hijacking and Spoofing

Error
Location
send_wechat.py:13
Finding

Unverified Process and Spoofable Window Targeting

Content
View full analysis
best_area: best_area = area best = (w, rect) except: pass return best[0] if best else None, best[1] if best else None ``` ### Technical Analysis The application trusts three hard-coded process identifiers and accepts the first process to which pywinauto can connect. Process identifiers are transient and reusable; they do not establish process identity. The code does not verify the executable name, executable path, digital signature, publisher, or any other WeChat-specific process property. After connecting, the code selects the largest visible window whose class contains `Qt51514QWindowIcon`. A generic Qt class and window area are not reliable application-authentication controls. An unrelated or attacker-controlled process can expose a matching window and receive the automated contact names and message content. The fallback PID `8` is especially inconsistent with the documented WeChat PID and further broadens the possibility of targeting an unrelated process. ### Attack Pa ...[truncated 1170 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
send_wechat.py:94
Finding

Message Sent to First Search Result Without Recipient Verification

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:16
Finding

Unpinned Third-Party Dependency Installation

Content
View full analysis
Remediation
View remediation
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 (4)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

Batch sending automates repeated message dispatch to multiple contacts without any interactive confirmation, recipient validation, rate limiting, or stop-on-error safety. In the context of a desktop UI automation tool for a personal messaging app, this materially increases the risk of spam, mass misdelivery, privacy breaches, and account sanctions if the wrong window or contact is targeted.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documentation promotes automated WeChat sending, including multi-recipient messaging, but does not prominently warn about misdelivery, accidental bulk sends, account restrictions, or the need for explicit recipient confirmation before transmission. In a UI-automation context, small targeting errors can send unintended messages to the wrong contact or many contacts at once, creating privacy, social-engineering, and spam risk.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The docstring at L045 says this function will send one message to a single contact, but the implementation never accepts or sends the actual message text. At L070 it types a hardcoded 'test message' or an empty string and returns before pressing Enter, which contradicts the documented intent of sending a real message.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The script sends a WeChat message immediately after selecting the first search result, with no confirmation, recipient verification, or dry-run mode. Because it drives the UI by focus and keystroke injection, any search mismatch, focus loss, or operator mistake can cause irreversible delivery to the wrong recipient.

Content

No source excerpt is available for this finding.