Security audit

内容过滤工具

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a content-filtering guide, but its activation text is mismatched and overbroad for an exec/write-capable skill.

Review this skill before installing. Its filtering functions are understandable, but the author should narrow the description and trigger conditions to explicit feed-filtering, rule-management, and blocked-content review requests, and clarify whether write access is actually required.

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 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 (7)

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The manifest presents the skill as a personal content-filtering tool, but the description broadens it into marketing copy generation, title optimization, and growth workflows. This capability mismatch can cause the agent to invoke the skill in unrelated contexts, leading to unauthorized scope expansion, misuse of exec-enabled behavior, and user deception about what the skill actually does.
Content: No source excerpt is available for this finding.

Vague Triggers

High

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The trigger description is overly broad and overlaps with common writing and content-creation requests, making accidental or inappropriate invocation likely. In an exec-capable skill, broad routing language raises the risk of the agent selecting this tool outside its intended scope and performing actions the user did not expect.
Content: No source excerpt is available for this finding.

Vague Triggers

High

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The activation condition is ambiguous and mismatched to the documented tool purpose, encouraging use during general marketing and growth tasks rather than filtering operations. This weakens tool-selection safeguards and can result in privilege-bearing tool access being granted in contexts where it is unnecessary.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The description explicitly states '支持中文交互', presenting a language constraint in the skill description without offering a user choice or documenting a justified locale limitation. Under the policy, forcing a specific language without opt-in is a natural-language policy concern.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The skill claims its core function is filtering, but the inline documentation advertises unrelated capabilities such as marketing copy and workflow support. This inconsistency increases the chance that an agent will treat the skill as a general-purpose content-generation tool, bypassing user expectations and safe tool-routing boundaries.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The core feature summary repeats that the skill provides Chinese interaction, again implying a fixed language mode rather than an opt-in choice. Because no alternative language option or justification is provided nearby, this appears to enforce a locale preference in the natural-language instructions.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The trigger conditions direct use of the skill for marketing promotion, ad placement, acquisition, and growth, which contradicts the stated filtering purpose. Overbroad and contradictory triggers are dangerous because they can cause inappropriate automatic activation in sensitive or unrelated workflows.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.