Back to skill

Security audit

Dataify Google News

Security checks for vulnerabilities and agentic risk

Overview

The advertised Google News search workflow is mostly coherent, but the package also includes under-disclosed broader scraping workflows and weaker API-token handling guidance.

Install only if you are comfortable giving the skill a Dataify API token and having requests sent to Dataify. Prefer a session-only DATAIFY_API_TOKEN, do not pass tokens on the command line, and avoid the persistent shell-profile setup shown in the Chinese guide unless you understand the plaintext secret risk. Treat the bundled non-news workflow scripts as outside the advertised Google News purpose.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/wait_for_task.py:33
Finding

API Token Exposed in Task Polling URL Query Strings

Content
View full analysis
") ``` It does not protect the request URL or any infrastructure logs created before the response is processed. This credential transfer is not required by the Skill’s declared Google News functionalit ...[truncated 1242 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/google_news.py:99
Finding

API Token Accepted Through Process Command-Line Arguments

Content
View full analysis
str | None: token = clean_value(token_arg) or clean_value(os.environ.get("DATAIFY_API_TOKEN")) if not token: return None if not token.lower().startswith("bearer "): token = f"Bearer {token}" os.environ["DATAIFY_API_TOKEN"] = token return token ``` ### Technical Analysis The Google News CLI permits users or calling processes to supply the API token using `--token`. Command-line arguments are not an appropriate channel for long-lived credentials because they may be exposed through: - Shell history files - Process inspection utilities - Operating-system process metadata - CI/CD job logs - Agent execution traces - Command auditing and endpoint monitoring - Diagnostic or crash reports Although the documented workflow generally recommends `DATAIFY_API_TOKEN`, the unsafe command-line interface remains active and can be used accidentally or by external automation. The function also writes the normalized bearer token back into `os.environ`. This does not persist it after process termination, but it unnecessarily broadens the token’s availability to subsequently invoked child processes. ### Attack Path 1. A user or automation system invokes the script as follows: ```bash python3 scripts/google_news.py --token SECRET_TOKEN --q "security news" ``` 2. The command is retained in shell history, CI logs, agent traces, or operating-system process metadata. 3. A local user, support operator, or compromised monitoring component reads the recorded command. 4. The observer extracts `SECRET_TOKEN`. 5. The token is reused to issue Dataify API req ...[truncated 491 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.zh-CN.md:21
Finding

Localized Documentation Recommends Persistent Plaintext Token Storage

Content
View full analysis
> ~/.bashrc source ~/.bashrc ``` ```bash echo 'export DATAIFY_API_TOKEN="your_token_here"' >> ~/.zshrc source ~/.zshrc ``` ### Technical Analysis These instructions cause a long-lived API token to be stored in plaintext in a user environment or shell startup file. Shell profiles are commonly: - Included in workstation backups - Collected in diagnostic and support bundles - Read by local processes running under the same account - Accidentally copied into dotfile repositories - Exposed by incorrect file permissions - Loaded into every future shell and inherited by child processes The guidance also conflicts with the account policy later in the same document, which says that session-scoped setup should be shown first and persistent setup should only be provided when requested. Persistent plaintext storage is not necessary for the Skill’s declared search functionality. A session-scoped environment variable is sufficient to authenticate a search process. ### Attack Path 1. A user follows the documented persistent setup command. 2. The API token is written into `.bashrc`, `.zshrc`, or the persistent Windows user environment. 3. The profile, registry-backed environment, backup, dotfile repository, or support archive becomes accessible to another party. 4. The party extracts the plaintext token. 5. The token is reused to perform Dataify API operations and consume the victim’s credits. ### Impact Assessment An attacker who reads the persistent configuration obtains the same Dataify API aut ...[truncated 374 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
Findings (29)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared description is narrow: search Google News for current news results only, and explicitly not for general Google web search. The supplied code does far more. It is not a dedicated Google News skill implementation; instead it is a shared workflow orchestrator for several business use cases. Depending on mode, it issues general Google Search queries, Google Shopping queries, Google News queries, and direct URL scraping requests. It also invokes Amazon comment and Google Maps review scrapers, extracts structured records, applies freshness/geography filters, scores leads, analyzes prices/reviews/brand mentions, and writes markdown/JSON reports. While one branch for brand monitoring does include a Google News action, that is only one small part of the actual behavior. The primary purpose and capabilities are materially broader and different from the declared description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared purpose says this skill searches Google News for current news results only. The supplied code does not implement Google News querying, news parsing, or even a fixed search workflow. Instead, it is a reusable submission wrapper for arbitrary catalog-defined Dataify tools: it reads a tool catalog, accepts any tool_sign, maps parameters, validates required fields, posts a job to a builder endpoint, and optionally waits for completion. This is a materially different and broader capability than the declared description. While a specific catalog entry might correspond to Google News elsewhere, this code chunk itself is generic infrastructure for many tools, so the description does not accurately represent the actual behavior shown.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The code does not implement Google News-specific search behavior. Its search function calls a generic SERP endpoint with parameters engine='google' and q=query, without any Google News/news-specific filter such as tbm=nws. That means it supports general Google search, directly contradicting the declaration 'Do not use for general Google web search.' Additionally, the code includes substantial undeclared capabilities unrelated to merely searching news: it can fetch and render arbitrary public URLs via the unlock function, clean/extract page content, and parse URLs from result payloads. These are materially broader capabilities than the declared purpose.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared purpose is an end-user search capability for Google News. The actual code chunk contains no search logic, no HTTP requests, no Google News interaction, and no result retrieval. Instead, it is a support/tooling script for inspecting or previewing request parameters by reading local markdown and Python files and printing a markdown table. This is a materially different primary purpose from the declared description, so it is a clear mismatch.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared description says the skill searches Google News for current news results. The supplied code contains no news-search logic, no Google News access, and no web search functionality. Instead, it implements generic infrastructure for parsing a task_id and polling/waiting for a remote task to finish via wait_for_task. This is a materially different primary purpose from the declared description, so it is a clear mismatch.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared purpose is news search via Google News, but the supplied code does not perform any search, network access, news retrieval, or Google-related functionality. Instead, its primary purpose is to provide platform-specific setup instructions for a DATAIFY_API_TOKEN environment variable and report whether it is configured. This is a materially different purpose and capability set from the description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared purpose says the skill searches Google News for current news results. The supplied code does not perform any Google News query or search operation. Instead, it waits on an existing Dataify scraper task by repeatedly calling Dataify status and download endpoints, handling authentication, credits, timeouts, and resume instructions. This is a materially different primary purpose and involves undeclared external resources (Dataify API/account) unrelated to the stated Google News search behavior.

Content

No source excerpt is available for this finding.

YARA rule 'backdoor_persistence': Backdoor persistence with malicious payloads (shell commands, SSH key injection, hidden root users) [malware]

High
Category
YARA Match
Confidence
75% confidence
Finding

YARA rule matched a known malware signature (reverse shell, backdoor, ransomware, C2 framework, or info stealer).

Content

Scanner excerpt · SKILL.zh-CN.md (reported line 40)May include surrounding context.

量,而不是只在当前终端临时设置。

Windows PowerShell,当前用户永久设置:

powershell
[Environment]::SetEnvironmentVariable("DATAIFY_API_TOKEN", "your_token_here", "User")

然后重新打开 PowerShell。如果当前会话也要立即生效,再执行:

powershell
$env:DATAIFY_API_TOKEN = "your_token_here"

macOS 或 Linux,bash 永久设置:

bash
echo 'export DATAIFY_API_TOKEN="your_token_here"' >> ~/.bashrc
source ~/.bashrc

macOS 或 Linux,zsh 永久设置:

bash
echo 'export DATAIFY_API_TOKEN="your_token_here"' >> ~/.zshrc
source ~/.zshrc

参考文件

  • references/ 目录下的 API 文档包含完整的字段列表和说明。
  • 当字段行为、允许的值或响应格式不明确时,请查阅参考文档。

注意事项

  • 始终使用 Content-Type: application/x-www-form-urlencoded 提交 API 请求。
  • 保持请求值为字符串类型。
  • 省略用户未请求的可选字段。

结果呈现

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The url_action logic routes user- or result-supplied URLs into a generic web unlocker and specialized review scrapers, including Amazon comments and Google Maps reviews. In the context of a Google News-only skill, this enables arbitrary external page fetching and specialized scraping that materially exceeds the declared purpose and increases privacy, compliance, and data-exfiltration risk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This shared workflow can dispatch not only Google News queries but also general Google Search, Google Shopping, arbitrary URL collection, and downstream detail scraping for multiple business-intelligence modes. For a skill explicitly described as 'Search Google News for current news results' and 'Do not use for general Google web search,' this creates a scope-expansion path that can be used to collect unrelated data and bypass the intended narrow capability boundary.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The manifest says this skill is for 'Search Google News for current news results' and explicitly says not to use it for general Google web search. However, the search implementation sends only engine=google and q parameters, with no news-specific selector such as a news vertical or tbm=nws, so it performs ordinary Google SERP queries rather than Google News-scoped queries.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
87% confidence
Finding

The skill instructs the agent to read environment variables, invoke a local Python script, access local reference files, and make outbound API requests, but the manifest declares no explicit tool scope or permissions. That mismatch weakens least-privilege controls and can allow broader execution capability than reviewers or policy engines expect.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger description includes broad everyday phrases such as news search/information requests, which can cause the skill to activate in situations the user did not specifically intend. In an agent setting, overbroad invocation increases the chance of unnecessary third-party API calls, unintended data disclosure to the external service, and unexpected credit consumption.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

L06 明确标注“Dataify Google News 中文版”,文件整体也仅以中文给出操作说明,但未说明这是可选语言、用户可切换,或该技能仅面向特定中文环境。根据规则,强制特定语言而无用户选择或正当限定,属于自然语言政策风险。

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 83)May include surrounding context.

md
## Account CTA policy

- Show a prominent Dataify account CTA only when the API token is missing, rejected/invalid, or the account has insufficient credits.
- For a missing token, offer https://dashboard.dataify.com/login?utm_source=skill and state: New accounts get 50 free credits, enough for about 6,000 trial results, valid for 7 days, and only successful requests are billed. Never ask the user to paste the token into chat.
- Detect the current operating system and shell. Show only the matching session-scoped setup command first (`export` for macOS/Linux shells, `$env:` for Windows PowerShell, or `set` for Windows Command Prompt). Show other platforms or persistent setup only when detection is ambiguous or the user asks.
- After the user says the token is configured, verify only whether `DATAIFY_API_TOKEN` is present; never print its value. If verification succeeds, continue the original task without asking the user to repeat it.
- Explain that persistent shell changes may require a new terminal or restarting the agent application. Do not recommend a project `.env` unless the execution path explicitly loads it, and ensure `.env` is ignored by version control.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.zh-CN.md (reported line 80)May include surrounding context.

md
## Account CTA policy

- Show a prominent Dataify account CTA only when the API token is missing, rejected/invalid, or the account has insufficient credits.
- For a missing token, offer https://dashboard.dataify.com/login?utm_source=skill and state: New accounts get 50 free credits, enough for about 6,000 trial results, valid for 7 days, and only successful requests are billed. Never ask the user to paste the token into chat.
- Detect the current operating system and shell. Show only the matching session-scoped setup command first (`export` for macOS/Linux shells, `$env:` for Windows PowerShell, or `set` for Windows Command Prompt). Show other platforms or persistent setup only when detection is ambiguous or the user asks.
- After the user says the token is configured, verify only whether `DATAIFY_API_TOKEN` is present; never print its value. If verification succeeds, continue the original task without asking the user to repeat it.
- Explain that persistent shell changes may require a new terminal or restarting the agent application. Do not recommend a project `.env` unless the execution path explicitly loads it, and ensure `.env` is ignored by version control.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

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

md
- Show a prominent Dataify account CTA only when the API token is missing, rejected/invalid, or the account has insufficient credits.
- For a missing token, offer https://dashboard.dataify.com/login?utm_source=skill and state: New accounts get 50 free credits, enough for about 6,000 trial results, valid for 7 days, and only successful requests are billed. Never ask the user to paste the token into chat.
- Detect the current operating system and shell. Show only the matching session-scoped setup command first (`export` for macOS/Linux shells, `$env:` for Windows PowerShell, or `set` for Windows Command Prompt). Show other platforms or persistent setup only when detection is ambiguous or the user asks.
- After the user says the token is configured, verify only whether `DATAIFY_API_TOKEN` is present; never print its value. If verification succeeds, continue the original task without asking the user to repeat it.
- Explain that persistent shell changes may require a new terminal or restarting the agent application. Do not recommend a project `.env` unless the execution path explicitly loads it, and ensure `.env` is ignored by version control.
- For an invalid token, direct the user to API-key management without implying that a new registration is required. For insufficient credits, direct the user to balance or recharge management.
- During normal submission, processing, and successful completion, do not promote registration or the Dashboard. Never expose the token or include it in CTA attribution parameters.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.zh-CN.md (reported line 82)May include surrounding context.

md
- Show a prominent Dataify account CTA only when the API token is missing, rejected/invalid, or the account has insufficient credits.
- For a missing token, offer https://dashboard.dataify.com/login?utm_source=skill and state: New accounts get 50 free credits, enough for about 6,000 trial results, valid for 7 days, and only successful requests are billed. Never ask the user to paste the token into chat.
- Detect the current operating system and shell. Show only the matching session-scoped setup command first (`export` for macOS/Linux shells, `$env:` for Windows PowerShell, or `set` for Windows Command Prompt). Show other platforms or persistent setup only when detection is ambiguous or the user asks.
- After the user says the token is configured, verify only whether `DATAIFY_API_TOKEN` is present; never print its value. If verification succeeds, continue the original task without asking the user to repeat it.
- Explain that persistent shell changes may require a new terminal or restarting the agent application. Do not recommend a project `.env` unless the execution path explicitly loads it, and ensure `.env` is ignored by version control.
- For an invalid token, direct the user to API-key management without implying that a new registration is required. For insufficient credits, direct the user to balance or recharge management.
- During normal submission, processing, and successful completion, do not promote registration or the Dashboard. Never expose the token or include it in CTA attribution parameters.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The file mixes English headings with Chinese field descriptions and specifically states 固定使用 for the engine, while the user-facing parameter guidance is presented in Chinese only. This creates a language/locale policy concern because the skill documentation effectively assumes a specific language without user opt-in or a documented region-specific justification.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module is explicitly described as shared business-workflow infrastructure rather than a narrowly scoped Google News skill. That mismatch is significant because it signals the file was reused without enforcing least privilege, making it easier for the packaged skill to expose broader search and scraping behavior than users and reviewers would expect.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The direct_request fallback can call general Google search, Google Shopping, and a generic webunlocker API when local scripts are unavailable. That means even if the surrounding repository is incomplete, this skill still retains broader-than-declared collection behavior and can access non-news scraping services, undermining the intended safety boundary of a Google News-only tool.

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/business_workflow.py (reported line 180)May include surrounding context.

python
def execute_action(action: dict[str, Any], token: str) -> subprocess.CompletedProcess[str]:
    invocation = command(action)
    if len(invocation) > 1 and Path(invocation[1]).exists():
        return subprocess.run(invocation, capture_output=True, text=True, encoding="utf-8", errors="replace", check=False)
    if action.get('capability', '').startswith('scraper-'):
        return subprocess.CompletedProcess([], 1, '', 'Required platform scraper is not installed; install it before executing this action.')
    return direct_request(action, token)

External Transmission

Medium
Category
Data Exfiltration
Confidence
70% 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 · scripts/catalog_builder.py (reported line 90)May include surrounding context.

python
def build_curl(tool, spider_parameters_json):
    return " \\\n".join([
        "curl -X POST '{}'".format(BUILDER_URL),
        "  -H 'Authorization: Bearer $DATAIFY_API_TOKEN'",
        "  -H 'Content-Type: application/x-www-form-urlencoded'",
        "  -d 'spider_name={}'".format(tool["spider_name"]),

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest describes a skill for searching Google News results. In this file, the code explicitly accesses the DATAIFY_API_TOKEN environment variable to authenticate outbound requests, which is a credential-handling capability not implied by the user-facing purpose of performing news searches.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The unlock function enables fetching arbitrary external URLs through a third-party web unlocking service, which exceeds the stated Google News search purpose and materially expands capability. In an agent-skill context, this can be abused to retrieve attacker-chosen content from arbitrary sites, including potentially sensitive internal or restricted destinations if upstream controls are insufficient, increasing SSRF-like and policy-bypass risk.

Content

No source excerpt is available for this finding.