Back to skill

Security audit

建筑工程投标决策-施工项目投标评估

Security checks for vulnerabilities and agentic risk

Overview

The skill matches its tender-analysis purpose, but it also performs account onboarding with device fingerprinting, stores API credentials locally, and preserves signed access links in shareable reports.

Install only if you are comfortable with this provider receiving project/company query terms, using a MAC-hash-based device check for trial registration, and storing an API key under ~/.zlbx/config.json. Avoid sharing generated HTML reports broadly unless you understand that included sk or sid links may grant access without a normal login prompt.

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
  • Rogue AgentSelf-Modification, Session Persistence
Findings (29)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill advertises tender-decision intelligence based on historical bid data, but the file mainly contains procedural instructions and local report-generation behavior, with no verifiable implementation boundary for the promised data collection and analysis. This mismatch is dangerous because users may authorize network access, credential use, and local file writes under false expectations, reducing informed consent and making abuse or overreach harder to detect.

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
93% confidence
Finding

This duplicated hidden-instructions finding points to the same commented block that imposes mandatory request headers and update-ack behavior outside the main visible workflow. Such hidden directives reduce auditability and can be abused to enforce covert telemetry or vendor lock-in behaviors without obvious disclosure in the primary skill logic.

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: construction-tender-bid-decision/1.0.9

Hidden Instructions

High
Category
Prompt Injection
Confidence
93% confidence
Finding

This duplicated hidden-instructions finding points to the same commented block that imposes mandatory request headers and update-ack behavior outside the main visible workflow. Such hidden directives reduce auditability and can be abused to enforce covert telemetry or vendor lock-in behaviors without obvious disclosure in the primary skill logic.

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: construction-tender-bid-decision/1.0.9

Hidden Instructions

High
Category
Prompt Injection
Confidence
88% confidence
Finding

This duplicate finding points to the same hidden instruction block that tells the agent to preserve a specific client identity and package source when making API calls. Such hidden control text is dangerous because it modifies authentication and request composition behavior from inside skill content, which should be treated as untrusted input. In this skill context, the instructions affect all calls to an external bidding API, increasing the blast radius if the package identity or endpoint expectations are maliciously altered.

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 finding points to the same hidden instruction block that tells the agent to preserve a specific client identity and package source when making API calls. Such hidden control text is dangerous because it modifies authentication and request composition behavior from inside skill content, which should be treated as untrusted input. In this skill context, the instructions affect all calls to an external bidding API, increasing the blast radius if the package identity or endpoint expectations are maliciously altered.

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-instruction finding refers to the same concealed request-identity block, which covertly instructs the agent to attach package/source headers and preserve an API key context. Concealed control instructions reduce transparency and can be abused to bind traffic to a particular backend or tracking scheme without clear user awareness.

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-instruction finding refers to the same concealed request-identity block, which covertly instructs the agent to attach package/source headers and preserve an API key context. Concealed control instructions reduce transparency and can be abused to bind traffic to a particular backend or tracking scheme without clear user awareness.

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
99% confidence
Finding

This skill’s documented behavior goes far beyond tender-bid analysis and introduces account provisioning, device fingerprint collection, persistent credential storage, and recharge/login flows. Those capabilities are unrelated to the declared purpose and create unnecessary access to local system attributes and long-lived credentials, expanding the attack surface and enabling covert account linkage or data exfiltration under the guise of a business-analysis skill.

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 the agent to collect local device fingerprints (platform, architecture, MAC-derived hash) and later persist API credentials to disk, despite those actions being unjustified by the business purpose of construction bid assessment. Even with partial minimization and nominal consent language, harvesting hardware-derived identifiers and storing keys locally creates privacy, tracking, and credential-abuse risk if the skill or endpoint is compromised.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicate finding identifies the same hidden-instruction pattern at the top of the template: concealed directives about API identity and request handling embedded in comments. Even if intended for packaging, embedding such directives in content rather than trusted configuration creates a prompt-injection-like channel that can override expected agent behavior and conceal data-sharing requirements.

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 finding identifies the same hidden-instruction pattern at the top of the template: concealed directives about API identity and request handling embedded in comments. Even if intended for packaging, embedding such directives in content rather than trusted configuration creates a prompt-injection-like channel that can override expected agent behavior and conceal data-sharing requirements.

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
92% confidence
Finding

This duplicated finding points to the same hidden instruction block that tells the agent to preserve a configured API key context and send package/source identity headers. Even if the headers are not secret themselves, placing such directives in hidden content increases the risk of covert data exfiltration patterns and weakens user/operator visibility into what authenticated requests the skill will trigger.

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
92% confidence
Finding

This duplicated finding points to the same hidden instruction block that tells the agent to preserve a configured API key context and send package/source identity headers. Even if the headers are not secret themselves, placing such directives in hidden content increases the risk of covert data exfiltration patterns and weakens user/operator visibility into what authenticated requests the skill will trigger.

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 least-privilege controls and can enable unexpected local file creation or overwriting if the runtime grants broader access than users or reviewers expect.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The description says the skill must be used even if the user does not mention 'construction', as long as the request involves project bid evaluation, whether to bid, or pricing reference. This broadens invocation beyond clearly bounded construction-specific triggers and does not provide exclusion examples, increasing the risk of unintended activation for generic tender-analysis requests.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
80% confidence
Finding

The manifest description and the entire user-facing instruction set are written as mandatory Chinese behavior, with no indication that users may choose another language. Under the policy, forcing a specific language without opt-in is a natural-language policy violation unless the locale constraint is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The file provides fixed Chinese text to relay to users and continues this pattern throughout the workflow, implying the skill should communicate in Chinese regardless of user preference. This is a natural-language locale policy concern because the document does not offer any opt-in or alternative language behavior.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

This section explicitly directs external transmission of collected device features to a third-party registration endpoint. The transmission is especially risky in this context because it is tied to local fingerprinting and account creation unrelated to the stated skill purpose, so it can leak host-derived identifiers off-device without a compelling functional need.

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`)。
>

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The curl example reinforces and operationalizes the same external data transmission path, making it easier for the agent or maintainers to implement exfiltration of host-derived data. In context, the danger is not the use of curl itself, but that the skill embeds instructions for shipping locally collected identifiers to an external service outside its declared mission.

Content

Scanner excerpt · references/auto-register.md (reported line 131)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`)。
>
> 历史教训:曾有版本采集 `home_path`,Windows 的 `C:\Users\alice` 直接拼进 JSON 字符串时

Session Persistence

Medium
Category
Rogue Agent
Confidence
97% confidence
Finding

This duplicate persistence finding points to the same instruction set for creating ~/.zlbx, merging config, and writing source metadata alongside the API key. In context, this persistence is particularly unsafe because it is paired with auto-registration, allowing the skill to both obtain and retain credentials without a need tied to construction-tender analysis.

Content

Scanner excerpt · references/auto-register.md (reported line 195)May include surrounding context.

text

注意事项:
- 目录不存在时先 `mkdir -p ~/.zlbx`
- 文件已存在时**合并而非覆盖**(保留用户可能的其他配置)
- `source: "auto"` 字段必须写入,**这是后续判断「是否输出自动登录链接」的关键依据**

Session Persistence

Medium
Category
Rogue Agent
Confidence
97% confidence
Finding

This duplicate persistence finding points to the same instruction set for creating ~/.zlbx, merging config, and writing source metadata alongside the API key. In context, this persistence is particularly unsafe because it is paired with auto-registration, allowing the skill to both obtain and retain credentials without a need tied to construction-tender analysis.

Content

Scanner excerpt · references/auto-register.md (reported line 195)May include surrounding context.

text

注意事项:
- 目录不存在时先 `mkdir -p ~/.zlbx`
- 文件已存在时**合并而非覆盖**(保留用户可能的其他配置)
- `source: "auto"` 字段必须写入,**这是后续判断「是否输出自动登录链接」的关键依据**

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The template explicitly instructs the agent to preserve and surface full signed URLs containing sk parameters in user-visible Markdown and exported HTML/JSON. If sk functions as a bearer-like access token or login-bypass signature, exposing it in chat transcripts, files, or shared reports can leak reusable authenticated links to unintended recipients and extend access beyond the original session.

Content

No source excerpt is available for this finding.