ClawHub
Back to skill

Security audit

Content Performance Analyst

Security checks for vulnerabilities and agentic risk

Overview

This is a content marketing reporting skill that reads specified business data and saves a monthly report, with no hidden execution or persistence found.

Use this only for authorized TCMS/content-marketing monthly reporting, and provide only the calendar, published-content, channel metrics, and product-map files needed for that report. Treat product-line maps and channel metrics as internal data, and consider narrowing the generic trigger phrases before publishing broadly.

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 (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger list includes very broad English phrases like "monthly report," "content review," and "performance analysis," which are common across many unrelated workflows. This can cause the skill to activate on generic user requests and be selected outside its intended TCMS-only context, increasing the risk of misrouting, unintended disclosure of brand-internal analysis behavior, or interference with other agent skills.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

L01 的“中文说明”表明该技能面向中文使用场景，但文件中未说明是否支持其他语言、是否需要用户显式选择中文，或为何必须固定为中文。按规则，若技能强制特定语言而没有用户选择或充分的场景性理由，属于自然语言政策风险。

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The trigger list includes broad everyday phrases such as '月度报告', '效果分析', '内容复盘', and English equivalents like 'monthly report' and 'performance analysis'. In an agent environment, overly generic activation phrases can cause accidental invocation outside the intended TCMS workflow, leading the skill to process unrelated user requests or activate in ambiguous contexts.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The read_when list contains broad phrases such as '效果分析', '数据复盘', 'monthly report', and 'performance analysis' that can match many unrelated user requests. This can cause the skill to activate outside its intended scope and read calendar, published-content, channel-effect, or private product-mapping data in contexts where the user did not specifically request this analysis, increasing the risk of unintended data exposure or inappropriate tool use.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The summary describes the skill's analytical functions but does not clearly warn that it consumes a private product map, even though the hard rules explicitly rely on private mapping data. This weak disclosure can mislead users or orchestrators about the sensitivity of the inputs and reduce opportunities to apply least-privilege controls or informed consent before the skill accesses internal business metadata.

Content

No source excerpt is available for this finding.