Back to skill

Security audit

招标商机发现-还没发标先知道

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its tender-opportunity purpose, but it also contains under-disclosed authenticated account, device-fingerprint, credential-persistence, and hidden request/acknowledgment behaviors that should be reviewed before installation.

Install only if you are comfortable with this vendor service receiving your search terms, using or creating a ZLBX API credential, storing that credential in your home directory, and collecting a hashed MAC-derived device identifier during auto-registration. Prefer preconfiguring ZLBX_API_KEY yourself and review the hidden request identity/update acknowledgment behavior before enabling it in sensitive environments.

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
98% confidence
Finding

The declared purpose is tender opportunity discovery, but the content also instructs local file generation, HTML export, possible PNG/PDF export, brand/link promotion, credential handling, and automatic registration flows that are not transparently reflected in the high-level description. This mismatch undermines user consent and policy enforcement because operators may approve the skill for data lookup while it also performs persistence, account lifecycle actions, and external navigation.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The activation text is overly broad and says the skill must be used even when the user does not explicitly ask for '商机', covering generic lead generation, market opportunity, and customer development requests. Overbroad routing increases the chance of unintended network queries, local writes, or account actions being triggered in contexts where a narrower or safer tool should have been selected.

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,对话清单输出后默认再用 `scripts/render_report.py` 生成一份可分享的 HTML 版商机清单并告知保存路径(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,对话清单输出后默认再用 `scripts/render_report.py` 生成一份可分享的 HTML 版商机清单并告知保存路径(详见 report-template.md「HTML 报告导出」)。

Hidden Instructions

High
Category
Prompt Injection
Confidence
89% confidence
Finding

This duplicate hidden-instruction finding points to the same concealed package-identity block, which instructs the agent to attach persistent package/source identifiers and perform a display acknowledgment back to the service. Such concealed control-plane behavior can facilitate tracking of agent usage and silently alter network behavior outside what a user would reasonably expect from a data-query skill.

Content

Scanner excerpt · SKILL.md (reported line 171)May include surrounding context.

  • 用户想持续跟踪 → 介绍订阅模式(见上方「订阅模式」)
  • 以上都不贴切 → 询问是否要对某条高价值商机深挖,或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: tender-opportunity-finder/1.0.8

Hidden Instructions

High
Category
Prompt Injection
Confidence
89% confidence
Finding

This duplicate hidden-instruction finding points to the same concealed package-identity block, which instructs the agent to attach persistent package/source identifiers and perform a display acknowledgment back to the service. Such concealed control-plane behavior can facilitate tracking of agent usage and silently alter network behavior outside what a user would reasonably expect from a data-query skill.

Content

Scanner excerpt · SKILL.md (reported line 171)May include surrounding context.

  • 用户想持续跟踪 → 介绍订阅模式(见上方「订阅模式」)
  • 以上都不贴切 → 询问是否要对某条高价值商机深挖,或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: tender-opportunity-finder/1.0.8

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicated finding reflects the same hidden instruction block telling the agent to preserve an API key context and send package identity headers. Even if intended for legitimate integration, placing such directives in skill content creates an injection channel where a model may be induced to exfiltrate environment-specific identifiers or follow network instructions from untrusted text.

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: opportunity-radar/1.0.3

Hidden Instructions

High
Category
Prompt Injection
Confidence
90% confidence
Finding

This duplicated finding reflects the same hidden instruction block telling the agent to preserve an API key context and send package identity headers. Even if intended for legitimate integration, placing such directives in skill content creates an injection channel where a model may be induced to exfiltrate environment-specific identifiers or follow network instructions from untrusted text.

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: opportunity-radar/1.0.3

Hidden Instructions

High
Category
Prompt Injection
Confidence
93% confidence
Finding

The file contains hidden operational instructions in comments directing the agent to attach specific identity headers and preserve package attribution. Hidden control instructions are dangerous because they influence agent behavior outside normal user-visible intent, and can be used to covertly steer requests, attribution, or downstream trust decisions.

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: opportunity-radar/1.0.3

Hidden Instructions

High
Category
Prompt Injection
Confidence
93% confidence
Finding

The file contains hidden operational instructions in comments directing the agent to attach specific identity headers and preserve package attribution. Hidden control instructions are dangerous because they influence agent behavior outside normal user-visible intent, and can be used to covertly steer requests, attribution, or downstream trust decisions.

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: opportunity-radar/1.0.3

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This file embeds an account lifecycle system—auto-registration, quota management, API-key persistence, and auto-login flows—inside a skill whose stated purpose is discovering tender opportunities. That materially expands the skill’s privileges and data-handling behavior beyond user expectations, creating risk of unauthorized account creation, credential handling, and opaque external service coupling.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill instructs collection of device fingerprint components (platform, architecture, MAC-derived hash) and local credential storage even though those capabilities are not necessary for a tender-opportunity finder’s core function. Even with hashing and stated consent language, this introduces tracking and persistent-auth behavior that is out of scope and privacy-sensitive in this context.

Content

No source excerpt is available for this finding.

Hidden Instructions

High
Category
Prompt Injection
Confidence
87% confidence
Finding

This duplicated hidden-instructions finding is also valid: the comment block contains non-user-visible directives governing API key use, client identity, and request metadata. Such hidden control text can be used to manipulate privileged behavior without transparent review, making the skill more dangerous than a plain reporting template.

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: opportunity-radar/1.0.3

Hidden Instructions

High
Category
Prompt Injection
Confidence
87% confidence
Finding

This duplicated hidden-instructions finding is also valid: the comment block contains non-user-visible directives governing API key use, client identity, and request metadata. Such hidden control text can be used to manipulate privileged behavior without transparent review, making the skill more dangerous than a plain reporting template.

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: opportunity-radar/1.0.3

Hidden Instructions

High
Category
Prompt Injection
Confidence
94% confidence
Finding

This duplicated finding reflects the same hidden-instructions issue: concealed prompt content tells the agent how to use an existing API key and which identity headers to send, including source attribution. Even if intended for routing or vendor analytics, placing these directives in hidden skill content creates a prompt-level mechanism for influencing authenticated outbound requests without clear user review.

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: opportunity-radar/1.0.3

Hidden Instructions

High
Category
Prompt Injection
Confidence
94% confidence
Finding

This duplicated finding reflects the same hidden-instructions issue: concealed prompt content tells the agent how to use an existing API key and which identity headers to send, including source attribution. Even if intended for routing or vendor analytics, placing these directives in hidden skill content creates a prompt-level mechanism for influencing authenticated outbound requests without clear user review.

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: opportunity-radar/1.0.3

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 64)May include surrounding context.

python
ROUTE_COLOR = {"拟建": "#7c5cbf", "意向": "#0b7a6a", "临期": "#b9770e"}

# 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 = "iVBORw0KGgoAAAANSUhEUgAAASsAAABgCAYAAABFTFSvAAAAAXNSR0IArs4c6QAAAERlWElmTU0AKgAAAAgAAYdpAAQAAAABAAAAGgAAAAAAA6ABAAMAAAABAAEAAKACAAQAAAABAAABK6ADAAQAAAABAAAAYAAAAACurZhBAAA2V0lEQVR4Ae2dB5xU1fXHzzZ6752ldxHpHTtqTKJJjEmMGkvKP/aIRk0UayzYxZJouokaYyzRiKCAoIB0kN47LB0WFnZh9//9vZmHb4eZnZndmWVW3vl8fjOv3fvuO+++3z3nvHvvM/PF14CvAV8DFUADafbM2OusKOMCKywoqgDlDV/ENIpemJ5vGYW5VmjbrChtlaXZYks7usRuumlb+ET+Vl8DvgYqkgYyzdJ6WqWskXa0IhU7UlnTzLgiK4K8jh7litK22FPPT4O03rb8zAk26hc5kVL6230N+BpIbQ1kWpEdsYJ8Hu4jqV3SeEuXlpZBkhaWmf49s/RvW9aRFfbs2D9ZweF/2i23bIo3O/94XwO+Bk6sBtJP7OmTeHZZV0JBgdmRI1mWnt7V0jIesozK7+H6/shGj66UxLP7Wfsa8DWQYA18fcnKq6iAWxggrbS0XpDWc1a34dP2zDMtvIf5y74GfA2krgZODrJy9e+SVuHROpae8ROC8i/bsy/1cHf7/74GfA2krgZOLrJy70NhoWJ0lXENz8Tces6eHHuqu8v/9zXgayA1NXBykpXuhayswkJeMKQPtHR7greGnVLzFvml8jXga0AaOHnJSlfvENbRLEtLG2BphaPtiSfqabMvvgZ8DaSeBk5ustL9EGEVFVW1jIwzLC3rlyzTWcsXXwO+BlJNAz5Z6Y4ohnXkaCPLSP++PfnCkFS7SX55fA34GjjZ3cDjakBatqUduQZ3sOpxu/wNvgZ8DZxQDfiWlav+gDtY3TKz+plVHupu9v99DfgaSA0N+GTlvQ8Bwmpm6YUX2htvaLiOL74GfA2kiAaST1bEr5Miycq3sLAm5e1jG7e3SUq5/Ux9DfgaKJUGkkNWRxgUfUSTHjAJQjqnKGRZ6wpklyQByybyEZpIQeC4jHQy1/HKV9sSIcovLS0NV7ARhcUd9MXXgK+BVNGAJlRJnEBGVbOy7Lz2be1bbdtax3p1LYtnf1tenk3euMneWrHSVu7cbTCNSAGyUU9yCILV9IxMq5KVaUfZdPioBh9DQDomTXxaZNUyM613s2Z2UYe2dlrDhlaD8+w5nG8zt22zd1atthlbmcZKZChyLKsUFtWxtKJTyOYfZc3KT+9rwNdAYjSQZk8//wKk8POyThGTAfF0b9DQHh022M5p3eq40slr27R/vz06a649N28uYSGzVnXr2vltWtvpLVpY29q1rHqlLDsC4ew8dMjm5+yw91avsbk5260mxHRr3952eZdOVp3lUNHxf1m01B6bNdu25uZChmUINzkEmXYIa+1dGPNKppPJCz2fv+5rwNdA+WsgIZZVBpbPKQ0b2avnn2tdsKaO4E6t3L3bpm/ZagcKjjgW1sAmTaxFzZr20JCB1q1BPQwXw0pqZw2rhu8lMBQr6mendLecgwctl2leOkFsynfdvn32xdYcZ3vzmjWsb+NG1rxGDefYJtWr2k2Tpth20pTJwioqghHTGllWVl1uiU9W5V8v/TP6GjhOAwkhq/pVqth9gwY4RLU3P98enTnHnsd62pN32IkrZeLi9Wve1P5w1unWtX59+1mP7scVJNyGLFw6EZGkEKKat32H3Txxkk1du94ho0xcQ+V3W7/e9oNOHe2c7NZ2Q699ds/n05nduJTCeZAM4m01sK4UbPfF14CvgRTQQJnJqjJhpcHNm9t5uHP
...[truncated 27 chars]

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding

The skill explicitly writes HTML reports to a local directory, but it does not declare any tool scope or allowed-tools boundary for file_write. Missing capability scoping weakens sandboxing and informed consent: an agent/runtime may permit broader file operations than users expect, especially since the skill also mandates automatic report export after execution.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The skill description and prescribed user-facing examples are entirely in Chinese and present a mandatory usage policy without offering any language choice or opt-in. Under the stated policy, forcing a specific language or locale without user choice is a reportable natural-language policy violation.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
87% confidence
Finding

This instruction documents transmission of collected device features and account-registration data to an external endpoint. The transmission itself is not proof of exploit code, but in this skill context it facilitates exfiltration of locally collected identifiers to a third-party service unrelated to the skill’s apparent business-search 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`)。
>

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

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
95% confidence
Finding

The skill directs persistence of an obtained API key in a user-home config file and immediate reuse in-session. Persisting credentials from within a content skill increases the blast radius of compromise, can surprise users, and creates long-lived authentication state unrelated to the skill’s core function.

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
95% confidence
Finding

The skill directs persistence of an obtained API key in a user-home config file and immediate reuse in-session. Persisting credentials from within a content skill increases the blast radius of compromise, can surprise users, and creates long-lived authentication state unrelated to the skill’s core function.

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**

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The file's title, required output template, and formatting instructions are all written to enforce Chinese-language output, including line L018's fixed Chinese report heading structure. There is no indication that the user may choose another language or that the skill is restricted to a China-specific locale context in a way that justifies mandatory Chinese output.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The template directs the agent to write a local JSON file and execute a local Python script as part of normal operation. That expands the skill from data retrieval/reporting into local code execution and filesystem modification, which increases attack surface and can be abused if report content, paths, or script behavior are influenced by untrusted data.

Content

No source excerpt is available for this finding.