ClawHub
Back to skill

Security audit

Financial Search Engine

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed financial search connector that uses a required Eastmoney API key and can save returned results locally by default.

Install only if you are comfortable giving this skill an Eastmoney EM_API_KEY and sending financial research queries to the configured Eastmoney service. Use --no-save for sensitive searches, rotate or revoke the key if it is exposed, and avoid placing confidential client or regulated information in queries unless that service is approved for it.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (8)

Tainted flow: 'req' from os.environ.get (line 176, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/get_data.py (reported line 187)May include surrounding context.

python
)

    try:
        with urllib_request.urlopen(req, timeout=timeout_seconds) as resp:
            raw_body = resp.read().decode("utf-8", errors="replace")
    except urllib_error.HTTPError as exc:
        err_body = exc.read().decode("utf-8", errors="replace") if exc.fp else ""

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The skill description overstates capabilities and omits material behaviors such as default local file writes, while the apparent implementation is only a thin wrapper over a single remote endpoint. This mismatch can mislead operators and users about data provenance, coverage, and side effects, causing inappropriate trust and unexpected persistence of potentially sensitive search content. Although not direct code execution, deceptive or inaccurate behavior declarations are a real security concern in agent ecosystems.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill advertises required environment access but does not declare an explicit tool scope for network and file-write behavior. That creates an authorization/transparency gap: a caller or review system may underestimate what the skill can do, while the implementation can access secrets, contact a remote service, and persist retrieved content locally. In this context, the risk is amplified because the skill handles an API key and writes query results by default.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The activation wording is very broad and can cause the skill to trigger on vague finance-related language without clear boundaries. In an agent setting, overly permissive triggering increases the chance of unnecessary remote queries, secret use, or unintended file output, especially when the tool performs network access and local persistence.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The module documentation explicitly states that runtime defaults are defined in code with 'no environment reads'. However, the implementation immediately reads os.environ.get("EM_API_KEY", "") and aborts if the variable is unset. This is a direct contradiction between the documented intent and actual behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code emits user-facing text in Chinese (默认输出目录为：) regardless of user preference. The file also contains additional Chinese-only operational/error/help text elsewhere, and there is no indication that the skill is region-specific or that users can choose a locale.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
75% confidence
Finding

The visible skill title and primary usage instructions are presented in Chinese, with no statement that the user may choose another language or locale for interaction/output. This can be a language policy concern if the organization requires language choice rather than a fixed default.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The skill saves retrieved financial search content to a predictable local workspace by default, even though persistence is not essential to the core search function. In an agent environment, automatic local writes can retain sensitive prompts, proprietary research, or regulated financial data longer than intended and expose it to other local processes, users, or later workflow steps.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.