Back to skill

Security audit

uos-cve-checker

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly does what it says, but its optional patch download mode can write remote-supplied files to unsafe paths without enough containment or verification.

Install only if you are comfortable with a CVE tool that contacts UnionTech and writes local output files. Avoid using -d in sensitive directories or automated environments until the downloader constrains paths to a dedicated folder, validates filenames and architectures, restricts download hosts, limits file size, and verifies patch authenticity. Treat the sudo dependency command as a manual prerequisite, not something to run automatically through an agent.

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 (1)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/cve_checker.py:136
Finding

Unvalidated Remote Download Metadata Enables Path Traversal and Arbitrary URL Fetching

Content
View full analysis
0: patch_item = patch_data["deb"][0] elif patch_data.get("oup") and len(patch_data["oup"]) > 0: patch_item = patch_data["oup"][0] # Handle patch as a list (old format) elif isinstance(patch_data, list) and len(patch_data) > 0: patch_item = patch_data[0] if patch_item: # Attempt to obtain url or download_url download_url = patch_item.get("url") or patch_item.get("download_url") download_pkg_name = patch_item.get("source") # If source is empty but download_url exists, extract the filename if not download_pkg_name and download_url: parsed = urlparse(download_url) path = parsed.path if path: filename = os.path.basename(path) if filename: download_pkg_name = filename else: download_pkg_name = f"patch_{arch}.bin" else: download_pkg_name = f"patch_{arch}.bin" if not download_url or not download_pkg_name: print(f"警告: 在 {pkg} 中缺少下载URL或包名,跳过下载。") continue # Create the architecture-specific directory ...[truncated 4718 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (8)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill documents capabilities that read inputs, write result files, and perform network requests, but it declares no explicit tool scope or permissions boundary. In an agent environment, missing scope metadata can cause overbroad execution or make reviewers and users unaware of what the skill is allowed to do, increasing the chance of unintended file access, disk writes, or external requests.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
76% confidence
Finding

The documentation instructs users to install a dependency with sudo apt, which normalizes privileged execution in the skill setup flow. While package installation may be legitimate, encouraging root-level commands increases risk if users run setup inappropriately, and in agent contexts any expectation of elevated privileges expands the blast radius of mistakes or future script changes.

Content

Scanner excerpt · SKILL.md (reported line 68)May include surrounding context.

依赖

bash
sudo apt install python3-requests

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/cve_checker.py (reported line 106)May include surrounding context.

python
}
        for attempt in range(retries):
            try:
                response = requests.post(url, headers=self.headers, json=cve_info_data, timeout=10)
                if response.status_code == 200:
                    data = response.json()
                    return data

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

When the optional download mode is enabled, the script automatically creates directories and writes patch files obtained from remote metadata to the local filesystem. Because the filename is derived from untrusted server-supplied data and there is no explicit trust prompt, path sanitization, or integrity verification, a compromised upstream service could cause unsafe file writes or silently place unexpected binaries on disk.

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/cve_checker.py (reported line 229)May include surrounding context.

python
}

                try:
                    response = requests.post(self.url, headers=self.headers, json=cvs_data)
                    response.raise_for_status()
                except requests.exceptions.RequestException as e:
                    print(f"查询 {cve_numb} 时请求失败: {e}")

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

SQP-3 applies to all file types and covers language or locale policy violations. The file consistently forces Chinese for the skill's user-facing description and usage instructions, with no indication that the user can opt into another language or that the skill is intentionally limited to a Chinese-speaking audience.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill states that it can automatically download patches and create output files, but it does not prominently warn that using -d triggers network activity and writes files to disk. This can lead to unexpected side effects in automated agent runs, especially in sensitive environments where outbound access or filesystem modification must be explicitly approved.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The request header hard-codes "Accept-Language: zh-CN,zh;q=0.9", which imposes a specific language/locale preference. This is a natural-language policy concern because the skill does not provide user opt-in or explain that requests will be forced to Chinese locale.

Content

No source excerpt is available for this finding.