Security audit

Novel Writing Assistant

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward novel-writing helper with local manuscript read/write behavior that fits its purpose, though it overstates some format-conversion support.

Install only if you want a novel-writing workflow that may read existing chapters and save manuscript or compiled text files in the project you point it at. Be aware that EPUB/PDF conversion is advertised but not actually provided by the bundled scripts, and the format checker is tailored to Chinese-style chapter markers.

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

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
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 (5)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 95% confidence
Finding: The code’s actual behavior is narrow and specific: it converts markdown chapter files into one plain text file by removing markdown syntax and concatenating chapter contents. While this partially fits the declared format-conversion aspect, it does not implement the broad primary purpose described as a comprehensive novel writing assistant. Most declared capabilities are absent, and even the conversion functionality is limited to .txt rather than the broader markdown-to-txt/epub/pdf claim. This is a material description-behavior mismatch, though there is no evidence of undeclared harmful or unrelated resource access.
Content: No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 89% confidence
Finding: The skill instructs reading existing chapters and saving files, which implies file read/write behavior, but it does not declare any tool scope or permissions boundaries. That can lead to over-broad filesystem access at runtime and makes it harder for users or the platform to understand what the skill may modify.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The description ends with a broad catch-all trigger ('any other novel writing tasks'), which can cause the skill to be invoked for loosely related requests beyond its validated behavior. In a skill that may read and write files, overbroad routing increases the chance of unintended file modification or use in contexts the user did not specifically authorize.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The skill tells the agent to read context and save chapters but does not warn users that project files may be modified. Without clear disclosure and confirmation, users may invoke the skill expecting advisory help while the agent performs persistent changes to local content.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The script's checks require Chinese-specific markers such as '第...章', '作者', '日期', and ending text like '结束|完', and the usage example names a Chinese-titled novel. This imposes a specific language/locale format in the skill behavior without offering a user choice or clearly documenting a justified region-specific constraint.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.