Back to skill

Security audit

废标风险检查-控标信号与限制条款识别

Security checks for vulnerabilities and agentic risk

Overview

This bid-analysis skill appears intended for its stated use, but it needs Review because it can create and store an API account and save reports containing signed access links.

Install only if you are comfortable sending bid, company, and project queries to Zhiliaobiaoxun, using or creating a ZLBX API key, and storing that key locally. Treat exported HTML reports as sensitive because they may include signed links that can open vendor resources for anyone who receives the report.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (33)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

声明描述的是一个分析型技能,其核心职责应是针对招标项目进行风险识别和投标决策分析,并基于全网历史数据产出评估结果。而这段代码只是将“已经生成好的 JSON 分析结果”渲染成 HTML 报告,带有样式、品牌 logo、引用附录、打印和导出长图/PDF 等展示功能。代码中没有任何网络访问、数据库查询、规则判断、模型推理、风险评分、控标信号检测或历史中标模式分析逻辑。因此其实际主用途与声明的主用途存在明显偏差:前者是报告展示器,后者是风险分析助手。虽然输入 JSON 字段名称与投标决策分析有关,但这只能说明它服务于该系统的结果呈现环节,不能等同于实现了声明中的分析能力。

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 81)May include surrounding context.

md
完整模式下,对话报告输出后默认再用 `scripts/render_report.py` 生成一份可分享的 HTML 版报告并告知保存路径(详见 report-template.md「HTML 报告导出」)。

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 85)May include surrounding context.

md
完整模式下,对话报告输出后默认再用 `scripts/render_report.py` 生成一份可分享的 HTML 版报告并告知保存路径(详见 report-template.md「HTML 报告导出」)。

Hidden Instructions

High
Category
Prompt Injection
Confidence
91% confidence
Finding

This duplicated finding points to the same hidden-instructions block that directs background header injection and update-ack behavior. Even if the content is operational rather than overtly malicious, concealing request-shaping logic in comments makes the skill harder to audit and increases the risk of covert telemetry or policy bypass in agent execution.

Content

Scanner excerpt · SKILL.md (reported line 163)May include surrounding context.

  • 报告里某个竞争对手威胁最大 → 引导用 zlbx-company-intel 对其深度背调
  • 以上都不贴切 → 报告涉及的企业完整档案与更多商机详情,引导访问知了商机大师 https://agent.zhiliaobiaoxun.com

安装包身份与升级检查:所有发往知了标讯 API 的请求(包括账户查询)必须额外携带以下两个请求头:

text
X-Skill-Package: bid-risk-redline-checker/1.0.9

Hidden Instructions

High
Category
Prompt Injection
Confidence
91% confidence
Finding

This duplicated finding points to the same hidden-instructions block that directs background header injection and update-ack behavior. Even if the content is operational rather than overtly malicious, concealing request-shaping logic in comments makes the skill harder to audit and increases the risk of covert telemetry or policy bypass in agent execution.

Content

Scanner excerpt · SKILL.md (reported line 163)May include surrounding context.

  • 报告里某个竞争对手威胁最大 → 引导用 zlbx-company-intel 对其深度背调
  • 以上都不贴切 → 报告涉及的企业完整档案与更多商机详情,引导访问知了商机大师 https://agent.zhiliaobiaoxun.com

安装包身份与升级检查:所有发往知了标讯 API 的请求(包括账户查询)必须额外携带以下两个请求头:

text
X-Skill-Package: bid-risk-redline-checker/1.0.9

Hidden Instructions

High
Category
Prompt Injection
Confidence
88% confidence
Finding

This duplicate hidden-instructions finding points to the same embedded directive block that attempts to control authenticated request identity through markdown comments. In the context of an agent skill, such concealed instructions increase the risk that the model will obey untrusted text to make credentialed requests under attacker-chosen package/source identifiers.

Content

Scanner excerpt · references/api-quick.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Hidden Instructions

High
Category
Prompt Injection
Confidence
88% confidence
Finding

This duplicate hidden-instructions finding points to the same embedded directive block that attempts to control authenticated request identity through markdown comments. In the context of an agent skill, such concealed instructions increase the risk that the model will obey untrusted text to make credentialed requests under attacker-chosen package/source identifiers.

Content

Scanner excerpt · references/api-quick.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Hidden Instructions

High
Category
Prompt Injection
Confidence
95% confidence
Finding

This duplicate hidden-instructions finding points to the same concealed request-identity block that dictates API headers and source attribution. Because it is embedded in comment syntax rather than transparent capability metadata, it can influence sensitive network behavior in a way that is not obvious to reviewers or users.

Content

Scanner excerpt · references/auto-register.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Hidden Instructions

High
Category
Prompt Injection
Confidence
95% confidence
Finding

This duplicate hidden-instructions finding points to the same concealed request-identity block that dictates API headers and source attribution. Because it is embedded in comment syntax rather than transparent capability metadata, it can influence sensitive network behavior in a way that is not obvious to reviewers or users.

Content

Scanner excerpt · references/auto-register.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The file turns a bid-risk analysis skill into an account bootstrap and device-fingerprinting workflow, including consent prompts, feature collection, registration, quota handling, and login flows. That behavior is outside the stated purpose of analyzing tenders and materially expands access to local state, network actions, and user/account manipulation, creating unnecessary privacy and security risk even if framed as operational support.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill instructs collection of platform, CPU architecture, and a hashed MAC address to deduplicate trial accounts. Even though the raw MAC is not transmitted, this is still device fingerprinting unrelated to bid review, and it creates a persistent identifier that can be used for tracking or account linkage beyond the user’s requested task.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The instructions tell the agent to create directories, merge JSON, persist API keys, and mutate local configuration for future sessions. For a reporting assistant, this is dangerous because it crosses from analysis into durable system modification and secret storage, which can change user environments and create long-lived trust or credential artifacts without a dedicated setup boundary.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicate hidden-instruction finding points to the same covert HTML-comment block that directs use of specific API identity headers. Even if operationally legitimate, placing such instructions in hidden prompt content creates an injection-like control channel and reduces auditability, making misuse or unnoticed credential-routing behavior more likely.

Content

Scanner excerpt · references/report-template.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicate hidden-instruction finding points to the same covert HTML-comment block that directs use of specific API identity headers. Even if operationally legitimate, placing such instructions in hidden prompt content creates an injection-like control channel and reduces auditability, making misuse or unnoticed credential-routing behavior more likely.

Content

Scanner excerpt · references/report-template.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Ssd 3

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The template explicitly requires preserving and exposing full API-returned URLs including the sk parameter, described as a login-bypass signature, in user-visible report content and exported JSON/HTML artifacts. This is a direct leakage of bearer-like signed access tokens that can be replayed by anyone who obtains the report, enabling unauthorized access to linked resources and durable sharing of privileged URLs.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
96% confidence
Finding

This duplicate finding points to the same concealed instruction block that tells the agent not to reuse another package's source/version and to send fixed package-identifying headers. In context, hidden directives about credentialed requests are especially risky because the skill is treated as untrusted content, and such instructions can manipulate agent behavior outside the user's visible intent while leaking installation/package provenance to the remote API.

Content

Scanner excerpt · references/workflow.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Hidden Instructions

High
Category
Prompt Injection
Confidence
96% confidence
Finding

This duplicate finding points to the same concealed instruction block that tells the agent not to reuse another package's source/version and to send fixed package-identifying headers. In context, hidden directives about credentialed requests are especially risky because the skill is treated as untrusted content, and such instructions can manipulate agent behavior outside the user's visible intent while leaking installation/package provenance to the remote API.

Content

Scanner excerpt · references/workflow.md (reported line 1)May include surrounding context.

请求身份:使用已配置的 API Key,并保留本包的 X-Client,携带下列完整身份头;不得沿用其它包的版本或来源。

text
X-Client: bid-decision/1.0.5

Obfuscated Code

High
Category
Supply Chain
Confidence
50% confidence
Finding

Code contains obfuscation (base64, hex encoding with execution). This is often used to hide malicious functionality.

Content

Scanner excerpt · scripts/render_report.py (reported line 71)May include surrounding context.

python
THREAT_COLOR = {"高": "#c0392b", "中": "#b9770e", "低": "#0d9463"}

# 知了标讯白色 logo(299x96 PNG base64 内嵌,保证报告离线/打印/转发时不裂图)
_LOGO_B64 = "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
...[truncated 27 chars]

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding

The skill explicitly states it will write HTML reports to a local directory, but it does not declare any tool scope or permissions boundary for file writing. Undeclared write capability weakens sandboxing and user visibility, and could allow broader-than-expected local file creation if the runtime grants generic file access.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The activation text is very broad and says the skill must be used even when users merely ask whether a bid has 'pitfalls' or whether it is worth bidding. Over-broad triggering can cause the agent to invoke external data access, account registration, and paid API consumption for ambiguous requests the user did not clearly intend to route through this skill.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The file presents the skill entirely in Chinese and instructs the agent to output a report in that language, but does not offer users any language or locale choice. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern unless the locale restriction is explicitly justified.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The document instructs the agent to send requests using a configured API key and to perform an account/balance check, but it provides no user-facing notice that project queries and account metadata will be transmitted to an external service. In an agent setting, this can lead to silent disclosure of user-supplied procurement targets and credential-backed account activity to a third party without informed consent.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown file contains prescriptive user-facing copy in Chinese and explicitly says it may be relayed verbatim to the user, which imposes a language choice on the interaction. The document does not offer an alternative locale or ask the user for language preference, so it conflicts with the language/locale policy criteria.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill directs the agent to inspect environment variables and local configuration files to obtain credentials before doing the requested analysis. Accessing local secrets and config state is a sensitive capability unrelated to the user’s substantive bid-risk question, and it broadens the chance of unauthorized credential use or accidental exposure.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
84% confidence
Finding

This finding reflects explicit instructions to serialize and transmit collected device and account data to an external registration endpoint. External transmission is expected for API use, but in this context it is sensitive because the transmitted data supports fingerprinting/account creation that is unrelated to the core bid-analysis purpose.

Content

Scanner excerpt · references/auto-register.md (reported line 130)May include surrounding context.

md
> ### ⚠️ 请求体必须用 JSON 序列化函数生成,不要手拼字符串
>
> 用 `json.dumps(payload)` / `requests.post(url, json=payload)` / `JSON.stringify(payload)`,
> 或 `curl -d @file`;**不要用字符串拼接,也不要用 Python 的 `str(dict)`**
> (后者产出单引号,服务端会报 `Expecting property name enclosed in double quotes`)。
>