Security audit

LINE Chat for Official Account

Security checks for vulnerabilities and agentic risk

Overview

This LINE skill mostly matches its purpose, but it can access and export customer chat content from a live business account without enough privacy controls.

Review before installing. Use it only for accounts where you are comfortable letting an agent operate a logged-in LINE OA browser session, read customer chats, send replies, and manage notes/tags. Add explicit confirmation before any reply or customer-image export, and prefer a skill-specific download folder instead of ~/Downloads.

A.I.G

Vulnerability Patterns

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
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (7)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 95% confidence
Finding: The declared description says the skill operates the LINE Official Account Manager via browser automation for message and chat management. However, this code chunk does not automate the LINE OA interface or perform any chat operations. It only assists initial setup by opening the LINE OA site, asking the user to paste the account chat URL, validating that URL, and writing it to a local configuration file. That is a materially different primary purpose from the declared operational/browser-automation behavior, so this is a mismatch.
Content: No source excerpt is available for this finding.

Ae1

High

Category: analysis-evasion
Confidence: 100% confidence
Finding: Referenced artifact was not completely inspected
Content: Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.
md
node scripts/setup.js

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The example trigger phrase "Check LINE messages" is broad and can easily match ordinary user requests that merely mention LINE, causing the skill to activate in situations the user may not have intended. Because this skill performs browser automation against a live messaging account, over-broad invocation increases the chance of unintended access to private customer chats or unintended actions in the LINE OA interface.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill uses shell execution to inspect ~/Downloads for downloaded chat images, which extends beyond browser-only LINE OA management into local filesystem access. This broadens the agent's privileges and creates risk of unintended exposure of unrelated local files or reuse of stale files from prior sessions.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill instructs the agent to download customer-supplied images from the LINE OA interface and then send those files to a user outside the application context. That creates a direct data exfiltration path for potentially sensitive customer content, with no access-control, authorization, or minimization checks described.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The instructions tell the agent to forward customer images to the user without any privacy notice, consent check, or policy boundary. In a messaging-support context, customer attachments may contain personal, financial, or other sensitive information, so silent forwarding materially increases privacy and compliance risk.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The script only clicks a button when its text includes the Chinese string "登入", which imposes a locale-specific assumption in the skill behavior. This can violate language/locale policy expectations because the skill does not offer opt-in, configuration, or alternative language handling.
Content: No source excerpt is available for this finding.

Static analysis

Detected: suspicious.dangerous_exec

Shell command execution detected (child_process).

Critical

Code: suspicious.dangerous_exec
Location: scripts/setup.js:30