Back to skill

Security audit

Novel Outliner - 大纲解析器

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward local novel-outline parser that generates prompt files, with the main caveat that it forces Chinese-language writing constraints and copies outline text into generated prompts.

Installers should understand that generated prompts are intentionally constrained to Chinese-only novel writing and may carry through any instruction-like text present in the source outline. Use trusted outlines or review generated prompts before sending them to an agent with sensitive tools or data access.

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
Findings (1)

T01 · Skill Instruction Hijacking

Warning
Location
scripts/parse_outline.py:44
Finding

Untrusted Outline Content Can Hijack Generated Agent Prompts

Content
View full analysis
Remediation
View remediation
{n}", f"{t}", f"{p}", ] ``` For stronger isolation, serialize the outline fields as JSON and supply them through a dedicated data channel supported by the downstream agent rather than concatenating them into its instruction text. ]]>
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (3)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill hard-codes a Chinese-only output constraint as an automatic injected requirement, overriding user preference and potentially causing policy or usability issues in multilingual contexts. While not a direct security exploit, hidden mandatory output constraints can mislead downstream agents and create unsafe prompt-shaping behavior by silently altering expected output requirements.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The CONSTRAINTS string requires '纯中文' and forbids Arabic numerals and English, which imposes a fixed language/locale policy. The file does not provide any user opt-in, configurability, or documented region-specific justification for this constraint.

Content

No source excerpt is available for this finding.