Back to skill

Security audit

获取微博热搜榜单,智能分类热点话题,生成分段式结构化日报。

Security checks for vulnerabilities and agentic risk

Overview

This skill is a narrowly scoped Weibo hot-topic report generator with some documentation and implementation caveats, but no evidence of hidden access, persistence, exfiltration, or destructive behavior.

Before installing, be aware that the skill contacts the disclosed third-party Weibo hot-list source and may include unfiltered hot-topic titles despite the README's filtering claim. The curl-based implementation could be tightened, but the observed behavior is proportionate to generating a live report.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T01 · Skill Instruction Hijacking

Note
Location
scripts/weibo_hot_daily.py:153
Finding

Mandatory Branded Footer Alters Agent-Generated Output

Content
View full analysis
Remediation
View remediation
str: # Generate report content... if include_attribution: lines.extend(["", "---", "by OpenClaw"]) return "\n".join(lines) ``` 4. Only enable attribution after the user explicitly requests it. 5. Add a test confirming that default report output contains only the requested hot-topic content and no unrelated branding. ]]>

T09 · Insecure Skill Coding Practices

Note
Location
README.md:10
Finding

Documented Political-Topic Filtering Is Not Implemented

Content
View full analysis
str: """分类热点""" # 财经科技 finance_keywords = [ 'A股', '股市', '指数', '股票', '经济', '公司', '企业', '科技', 'AI', '互联网', '芯片', '新能源', '汽车', '股价', '涨停', '跌停', '小米', '华为', '苹果', '特斯拉', '茅台', '比特币', '基金', '月薪', '薪资', '裁员', '招聘', '就业', '房价' ] for kw in finance_keywords: if kw in title: return '财经科技' # 社会民生 society_keywords = [ '警方', '事故', '安全', '死亡', '受伤', '法院', '判决', '医院', '教育', '学校', '老师', '学生', '家长', '孩子', '幼儿园', '高中', '大学', '博士', '诈骗', '盗窃', '犯罪', '火灾', '地震', '暴雨', '台风', '灾害', '救援' ] for kw in society_keywords: if kw in title: return '社会民生' # 文体娱乐 entertainment_keywords = [ '电影', '电视剧', '综艺', '明星', '演员', '歌手', '导演', '国足', '足球', '篮球', 'NBA', '比赛', '冠军', '演唱会', '专辑', '歌曲', 'MV', '票房', '定档', '官宣', '恋情', '结婚', '离婚', '绯闻', '塌房', 'cp', '粉丝' ] for kw in entertainment_keywords: if kw in title: return '文体娱乐' return '网络热点' ``` ### Technical Analysis The README states that the Skill filters politically sensitive topics, but the implementation contains no political-content filtering stage. `categorize_hot()` only assigns each title to one of four categories. Titles that match none of the listed keywords are assigned to the internet-trends category and remain eligible for ou ...[truncated 1305 chars]
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (7)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

“使用方法”只给出几个示例短语,没有定义正式触发列表,也没有说明必须满足的上下文、参数或排除条件。对于 markdown 技能说明,这种缺少边界的触发描述属于模糊激活条件,可能导致意外调用。

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

README 的使用方法将“今天的微博热搜”列为触发示例,但未说明这是仅在特定技能选择上下文中生效,还是任意提到该短语都会触发。该表述接近日常自然请求,且文档未提供触发边界、限定条件或反例,容易造成技能被过宽匹配。

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill declares executable dependencies (python3, curl) and references a helper script, but does not define any explicit tool scope such as permissions or allowed-tools. That creates an authorization gap where the runtime may permit broader shell/network actions than are necessary, increasing the chance of unintended command execution or unbounded outbound requests to third-party infrastructure.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file defines invocation examples such as “今天的微博热搜” and “生成微博热点日报” without clarifying boundaries or exclusion conditions. Those phrases are generic enough that a normal request for information about Weibo trends could unintentionally invoke the skill rather than a broader conversational response.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/weibo_hot_daily.py (reported line 54)May include surrounding context.

python
def fetch_weibo_hot(url: str = "https://weibo.g.renyuzhuo.cn/") -> list:
    """获取微博热搜数据"""
    try:
        result = subprocess.run(
            ['curl', '-s', '-m', '15', url],
            capture_output=True,
            text=True,

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This skill performs data retrieval by spawning an external curl process, which grants broader command-execution capability than the task requires. In an agent/skill environment, that extra capability is risky because future changes, parameterization of url, or hostile runtime conditions could turn a harmless fetch into SSRF, policy bypass, or abuse of the local execution environment.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

该 README 全文以中文定义技能名称、调用示例和输出格式,未说明是否支持其他语言或允许用户选择输出语言。按照语言/区域策略,若技能实际强制单一语言而未提供选择或明确合理的限定,属于自然语言层面的策略问题。

Content

No source excerpt is available for this finding.