ClawHub
Back to skill

Security audit

Search Web - CN

Security checks for vulnerabilities and agentic risk

Overview

The skill is a plausible Volcengine web-search helper, but it under-discloses that it sends full user queries and a bearer credential to an external API and can silently fall back to a broader Ark API key.

Review before installing. Use a dedicated, low-privilege VOLCENGINE_SEARCH_API_KEY, avoid exposing a general VOLCENGINE_ARK_API_KEY in the runtime environment, and do not send secrets, private source code, internal URLs, or regulated data as search queries unless this provider and endpoint are approved for that data.

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
Findings (1)

other

Error
Location
scripts/volcengine_search_web.py:18
Finding

API Credential and User Query Disclosure to an Unrelated Third-Party Endpoint

Content
View full analysis
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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 (9)

Tainted flow: 'headers' from os.getenv (line 45, credential/environment) → requests.post (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/volcengine_search_web.py (reported line 65)May include surrounding context.

python
if stream:
            # Streaming response
            print("\n=== Response (Streaming) ===\n")
            response = requests.post(API_URL, headers=headers, json=payload, stream=True)
            response.raise_for_status()

            full_content = ""

Tainted flow: 'headers' from os.getenv (line 45, credential/environment) → requests.post (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/volcengine_search_web.py (reported line 108)May include surrounding context.

python
print("\n")
        else:
            # Non-streaming response
            response = requests.post(API_URL, headers=headers, json=payload)
            response.raise_for_status()
            data = response.json()

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

There is a material mismatch between the declared behavior and the detected implementation details, including use of an unexpected external endpoint, undeclared credential requirements, and a CLI-style execution path not reflected in the skill metadata. Behavior/description mismatches are dangerous because they can conceal data exfiltration destinations, bypass reviewer expectations, and cause users to disclose sensitive queries or credentials to an unanticipated third party.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The README clearly promotes a networked search skill backed by Volcengine's API, but it does not explicitly warn users that their prompts and possibly related metadata will be sent to a third-party service for processing. In an AI coding/tooling context, users may submit sensitive source code, internal URLs, tokens, or proprietary questions, so missing disclosure materially increases the risk of unintended data exposure.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill appears to use both environment-sourced secrets and outbound network access, but the manifest does not declare any tool scope or permissions. This weakens transparency and policy enforcement because users and hosting platforms cannot clearly evaluate that the skill will read credentials and transmit data externally before invocation.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% 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/volcengine_search_web.py (reported line 65)May include surrounding context.

python
if stream:
            # Streaming response
            print("\n=== Response (Streaming) ===\n")
            response = requests.post(API_URL, headers=headers, json=payload, stream=True)
            response.raise_for_status()

            full_content = ""

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% 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/volcengine_search_web.py (reported line 108)May include surrounding context.

python
if stream:
            # Streaming response
            print("\n=== Response (Streaming) ===\n")
            response = requests.post(API_URL, headers=headers, json=payload, stream=True)
            response.raise_for_status()

            full_content = ""

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The script forwards the user's raw question content to an external third-party API without any explicit notice, consent flow, or filtering for sensitive data. In an agent-skill context, users may reasonably assume local handling, so secrets, personal data, internal URLs, or proprietary prompts can be unintentionally disclosed to the remote provider.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill documentation describes search functionality but does not clearly warn that user prompts and the returned content flow through a third-party web search/Q&A provider. This is a privacy and trust issue because users may unknowingly send proprietary code, internal questions, or regulated data to an external service with separate retention and handling policies.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.