Back to skill

Security audit

Goal Manager - 目标管理

Security checks for vulnerabilities and agentic risk

Overview

This is a local Chinese-language goal tracker that writes goal data to JSON files and has no evidence of network access, credential use, hidden execution, or exfiltration.

Install only if you are comfortable with a Chinese-language local goal tracker storing goals in files beside the skill. Be careful with delete or cancel requests because matching goals may be removed without a confirmation step.

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
Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (8)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The README is written entirely in Chinese and provides examples only in that language, without indicating multilingual support or giving users a language choice. This can lead to accessibility and usability issues, causing users or downstream systems to assume a fixed language mode and mishandle requests from users in other languages.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The example triggers are generic everyday phrases such as setting goals, updating progress, and doing a monthly review. Because they closely overlap with normal user requests, the skill may activate unintentionally or be selected in contexts broader than intended, which can cause misrouting, confusion, or inappropriate handling of user input.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The entire skill interface, trigger phrases, commands, and examples are specified only in Chinese, with no indication that users may choose another language or opt in to a locale-specific experience. This can violate language/locale policy when a skill implicitly forces one language without user choice or documented justification.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill advertises a destructive 删除 [目标名] command but provides no confirmation step, undo mechanism, or warning about permanent data loss. In a goal-tracking skill that stores user data in local files, ambiguous or accidental invocation could erase records and reviews unexpectedly.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The function save_reviews is documented as saving review data, and the module advertises review/tracking features, but no code path calls save_reviews and generate_review only returns a formatted string without storing any review content. This creates an intent-code divergence where the code suggests persisted review management exists, but actual behavior only generates a template.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The delete path removes goal data immediately based on a natural-language query substring match and persists the change without confirmation, preview, or undo. In an agent setting, ambiguous phrasing, prompt injection, or accidental invocation could cause irreversible loss of user data, especially because deletion criteria are broad and silent.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The module description, templates, command examples, and all user-facing responses are written in Chinese, effectively forcing a specific language for interaction. The file does not provide an opt-in language selection mechanism or a documented justification for the locale restriction.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The comment 确定存储位置 implies goals are stored according to their duration category, but the implementation places all non-yearly and non-quarterly goals, including half_year, into monthly_goals. This is a direct mismatch between the documented intent of duration-based storage and the actual categorization logic.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.