Security audit

Web Novel Creator (网文创作协调层)

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed web-novel project coordination helper that moves chapter/planning files and delegates memory index updates, with no evidence of hidden code, credential use, remote payloads, or exfiltration.

Install only if you want a Chinese web-novel workflow helper that may move or rename chapter outputs, create planning/title files, and ask Memory Manager Pro to update persistent project indexes. Before using it on an existing workspace, confirm the target novel project path and review any proposed migrations or renames, especially if other skills also depend on the same files.

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

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The README states that the skill automatically detects external skills and migrates their outputs into a shared project directory, but it does not disclose the risk of cross-skill file renaming or relocation. Because this skill is specifically designed as a compatibility layer across multiple tools, silent normalization of files is more dangerous here than in a single-purpose skill: it can disrupt assumptions made by other skills and affect a larger set of project data.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The README explicitly advertises automatic migration and full index updating, but it does not warn users that invoking the skill may modify project files and shared indexes. In a shared project workspace, undocumented write operations can lead to unintended file moves, overwrites, or state changes that users did not consent to, especially when multiple skills rely on the same directory structure.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The manifest description is entirely in Chinese and does not indicate that users may interact in other languages or choose their preferred locale. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The operational instructions throughout the markdown are presented only in Chinese, which can impose a fixed language behavior on users. Because the file does not offer a language/locale choice or explain a necessary region-specific limitation, this matches the language policy violation criteria.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: Line L183 says the skill '不直接创作正文', yet the flow at L195 explicitly permits '直接创作（fallback）'. The same sentence also carves out planning generation as an exception, so the documentation internally contradicts itself about whether the skill is strictly non-creative.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The manifest repeatedly frames web-novel-creator as a compatibility/coordinator layer that should not replace dedicated writing skills. However, the documented rule flow explicitly allows '直接创作（fallback）' and states it may generate follow-up planning when none exists, which goes beyond pure coordination and archival behavior.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The example trigger phrases include broad, natural requests such as '继续创作天道养殖场', '查看当前任务', and '查一下之前的编码任务'. In an agent skill ecosystem, these are realistic everyday utterances that may unintentionally activate this skill when the user intended a different tool, causing context bleed, misrouting, or unauthorized access to project/task/index data.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.