Security audit

Novel Writer

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent novel-writing helper with disclosed templates and local helper scripts, and the only notable issue is broad trigger wording that could invoke it during general writing requests.

Install only if you want a general-purpose Chinese novel-writing assistant. Be aware that short prompts like '润色' or '续写' may route ordinary writing tasks into this skill, and only run the helper scripts on manuscript files or directories you intentionally choose.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (8)

Vague Triggers

High

Confidence: 95% confidence
Finding: The advertised trigger phrases are extremely broad generic terms like “续写”, “润色”, and “逻辑检查”, which are common in ordinary conversation and can cause unintended auto-invocation of the skill. In an agent environment, this increases the chance that the skill activates on unrelated user requests, expanding access to user content or causing unexpected behavior without clear user intent.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The trigger phrases for story ideation are very broad and overlap with ordinary user requests such as '写小说' or '创作故事'. In a skill-routing system, this can cause unintended activation, leading the agent to apply this skill in contexts where the user did not explicitly intend to invoke it, which may interfere with normal behavior or route sensitive text into the wrong workflow.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: Continuation triggers like '续写', '扩写', and '继续写' are highly generic and likely to appear in many normal writing interactions. Without constraints, the skill may be invoked accidentally during unrelated editing or drafting tasks, causing prompt-routing confusion and unpredictable outputs.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: Polishing/editing phrases such as '润色', '修改文章', and '改写' are common across many general-purpose writing tasks, not just novel-writing. This increases the chance of overmatching and accidental invocation, which can misapply this skill outside its intended scope and disrupt other tools or workflows.

Vague Triggers

Medium

Confidence: 96% confidence
Finding: The top-level trigger list repeats short, generic phrases without any disambiguation, reinforcing accidental activation risk across multiple common writing scenarios. Because these phrases are displayed as the main entry points, they make unsafe routing behavior more likely at the skill-discovery layer.

Vague Triggers

Medium

Confidence: 74% confidence
Finding: The trigger phrases for story creation are broad enough to match common conversational requests, increasing the chance of unintended skill activation. In an agent ecosystem, overbroad activation can cause the wrong skill to take over context, leading to unexpected file access, incorrect workflow routing, or disclosure of user-provided content to components the user did not intend to invoke.

Vague Triggers

Medium

Confidence: 76% confidence
Finding: Generic continuation phrases like 'continue writing' or 'what happens next' are highly collision-prone with normal assistant use. This can cause accidental activation during unrelated chats, which is especially risky because the skill also references multi-file reading and content-processing workflows that may be triggered without clear user intent.

Vague Triggers

Medium

Confidence: 78% confidence
Finding: Editing triggers such as broad 'polish' or 'rewrite' style requests overlap heavily with routine assistant tasks, making unintended invocation plausible. Misrouting editing requests to a specialized skill can expose user drafts to unnecessary processing paths and create confusion about what transformations, retention, or file operations may occur.

Static analysis

No suspicious patterns detected.