Security audit

年度增长计划

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent growth-planning assistant, but it can create or overwrite local planning files after broad triggers without clear confirmation or containment.

Install only if you are comfortable with the agent creating and updating local growth-plan documents. Before using it, give an explicit target filename and ask the agent to confirm before writing or overwriting anything, especially if the plan contains income, customer, budget, or business strategy details.

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
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 (4)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger list includes very generic phrases such as “OKR”, “复盘”, “回顾”, and “我的目标”, which are common in ordinary conversations and can cause the skill to activate outside the user’s intended context. This creates a prompt-scope issue where the agent may apply file-reading/writing workflow and growth-planning behaviors to unrelated requests, increasing the chance of unintended actions or inappropriate context takeover.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill explicitly instructs the agent to generate and write a Markdown file but does not require notifying the user that a local file will be created. Silent file creation is dangerous because it crosses from conversation into persistent side effects, which can surprise users, clutter workspaces, or create sensitive planning documents in unintended locations.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: Review mode says the plan will be saved by overwriting the existing file, but the skill does not require a warning or confirmation before destructive modification. Overwrite behavior is riskier than initial creation because it can permanently replace prior user content, including historical notes or manually edited data, if the wrong file is selected or the update is mistaken.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The mode-selection rule sends anything not containing “回顾/复盘/更新” into plan-creation mode, which is an ambiguous default for a broad set of user inputs. In practice, this can misclassify requests and lead the agent to create or modify planning artifacts when the user only wanted discussion, advice, or analysis, making downstream file operations more likely without clear intent.

Static analysis

No suspicious patterns detected.