Back to skill

Security audit

jiege-openclaw-video

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent video generator, but it can send an arbitrary configured OpenClaw API key to a fixed third-party service and open returned URLs without confirmation.

Review before installing. Only use this skill with a dedicated, low-scope API key intended for the WanjieData video service, and avoid using it in an OpenClaw profile that contains unrelated provider keys. The skill should be fixed to select an explicit provider, disclose the endpoint and data sent, require confirmation before opening returned links, validate result URLs, and provide job status and cancellation controls.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (2)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
model/scripts/veo_worker.py:19
Finding

Arbitrary OpenClaw Provider Credential Transmitted to a Hard-Coded Third-Party Service

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
model/scripts/veo_worker.py:49
Finding

Untrusted URL from Remote API Response Automatically Opened in the User's Browser

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
Findings (17)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The script automatically opens a URL returned from remote model output using os.startfile, which creates a local execution-side effect based on untrusted external content. A malicious or compromised upstream service could cause the host to open phishing pages, dangerous file/URI handlers, or other unintended resources without user confirmation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The README explicitly states that the skill will automatically open the system's default browser after task completion, with no mention of user confirmation or safety checks. Automatically launching a browser from an agent workflow can be abused to trigger untrusted URLs, create unexpected external interactions, or bypass normal user review of outbound actions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill only activates on the Chinese prefix '生成视频:' and responds in Chinese, with no indication that other languages are supported or that the user can choose a locale. This is a natural-language policy concern because the behavior enforces a specific language without documented opt-in or justification.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The hook launches a detached background Python subprocess based on a chat trigger, with no user confirmation, job controls, rate limiting, or visibility into what the worker will do. Even though arguments are passed safely to spawn without shell interpolation, this still enables unbounded background execution from user input and makes abuse, resource exhaustion, and opaque side effects more likely.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The script reads an API key from the user's home-directory configuration file without any validation, disclosure, or clear justification beyond convenience. Accessing local credentials inside a skill increases the trust boundary and can lead to unauthorized use of secrets, especially when those secrets are immediately sent in an Authorization header to an external service.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The script silently reads API credentials from a user home-directory config file without warning, which is a privacy and consent problem even if the key is used only for the apparent service call. Hidden credential access is risky in a skill context because users may not expect local secret discovery as part of normal operation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The script sends the user-provided prompt and runtime parameters to an external API without any explicit notice that the data will leave the local machine. In a skill environment, undisclosed remote transmission can expose sensitive prompt contents and metadata to third parties.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This duplicate finding reflects the same outbound POST request to an external service carrying potentially sensitive prompt content and authentication material. In the context of a skill with no stated metadata or data-sharing notice, that transmission is more dangerous because users cannot assess or approve the privacy implications.

Content

Scanner excerpt · model/scripts/veo_worker.py (reported line 31)May include surrounding context.

python
}
        
        print(f"[*] 提交任务: {model}, 时长: {seconds}s")
        res = requests.post("https://maas-openapi.wanjiedata.com/api/v1/chat/completions", 
            headers={"Authorization": f"Bearer {key}", "Content-Type": "application/json"},
            json=payload, stream=True)

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

This duplicate finding reflects the same outbound POST request to an external service carrying potentially sensitive prompt content and authentication material. In the context of a skill with no stated metadata or data-sharing notice, that transmission is more dangerous because users cannot assess or approve the privacy implications.

Content

Scanner excerpt · model/scripts/veo_worker.py (reported line 31)May include surrounding context.

python
}
        
        print(f"[*] 提交任务: {model}, 时长: {seconds}s")
        res = requests.post("https://maas-openapi.wanjiedata.com/api/v1/chat/completions", 
            headers={"Authorization": f"Bearer {key}", "Content-Type": "application/json"},
            json=payload, stream=True)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Automatically opening a returned URL without warning or confirmation removes user control over a potentially unsafe action initiated by untrusted remote output. This can expose users to phishing, malicious downloads, or abuse of local URI handlers, making the behavior more dangerous in an agent skill than in a purely interactive application.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
90% confidence
Finding

This function launches a separate Python process using user-influenced arguments (prompt, model, seconds) with no validation, restriction, or lifecycle control. Although shell=True is not used, arbitrary subprocess spawning from a skill can still be abused for unauthorized code execution paths, resource exhaustion, or detached background activity that is difficult for the caller to monitor or stop.

Content

Scanner excerpt · model/scripts/video_interface.py (reported line 18)May include surrounding context.

python
]
    
    # 异步启动进程,与 OpenClaw 进程完全解耦
    subprocess.Popen(cmd, creationflags=subprocess.CREATE_NEW_PROCESS_GROUP)
    return f"[*] 任务已提交: {prompt}"

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The code starts a background subprocess immediately and returns a success message without any user-facing warning, approval step, or indication that a detached process will continue running independently. In an agent skill context, this reduces transparency and makes it easier to trigger unexpected local execution or persistent resource consumption without informed consent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill documentation states that intercepted prompts automatically trigger backend video generation and will open a browser page on completion, but this behavior is not clearly presented as a user-facing warning. Automatic execution plus unsolicited browser launching can surprise users, cause unintended API usage or charges, and create opportunities for unsafe workflow triggering if users do not understand the side effects of normal chat input.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The file includes Chinese-only comments and runtime messages such as status output, but offers no language selection or explanation that the skill is intentionally locale-specific. This creates a natural-language policy concern because it forces a specific language experience without opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The file's human-facing comments and returned status message are written only in Chinese, with no indication that language selection is optional or configurable. This can violate language/locale policy when a skill implicitly forces one language without offering user choice.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
95% confidence
Finding

The dependency is specified as requests>=2.25.1, which allows installation of many later versions without any upper bound or exact pin. This harms reproducibility and can unintentionally introduce vulnerable or breaking releases during future installs, making supply-chain risk harder to assess and control.

Content

Scanner excerpt · requirements.txt (reported line 1)May include surrounding context.

text
requests>=2.25.1

Unverifiable Dependency: requests has 16 known advisory(ies) (CVE-2014-1830 (Exposure of Sensitive Information to an Unauthorized Actor in Requests); CVE-2024-47081 (Requests vulnerable to .netrc credentials leak via malicious URLs); CVE-2024-35195 (Requests `Session` object does not verify requests after making first request wi) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
88% confidence
Finding

The manifest references requests without pinning an exact version, while the package has multiple known advisories across different releases. Because the installed version is unconstrained beyond a minimum, it is not possible to verify from this file alone whether deployment will pull a vulnerable release, creating avoidable uncertainty in the dependency security posture.

Content

No source excerpt is available for this finding.