Back to skill

Security audit

Sysinfo Tool

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent system-information tool, but it collects full process command lines by default without clearly telling users, which can expose sensitive local details.

Review before installing or using in shared logs, support bundles, CI, or agent transcripts. Prefer brief or section-specific output when possible, and avoid JSON/default reports on machines where command-line arguments may contain tokens, passwords, internal paths, or customer data.

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

T05 · Unauthorized Access and Privilege Escalation

Note
Location
scripts/sysinfo.py:144
Finding

Undisclosed System-Wide Process Reconnaissance Enabled by Default

Content
View full analysis
List[Dict[str, Any]]: """Get top processes.""" processes = [] try: result = subprocess.run( ['ps', 'aux', '--sort=-pcpu'], capture_output=True, text=True ) lines = result.stdout.splitlines()[1:11] # Top 10 for line in lines: parts = line.split(None, 10) if len(parts) >= 11: processes.append({ 'user': parts[0], 'pid': parts[1], 'cpu': float(parts[2]), 'mem': float(parts[3]), 'command': parts[10] }) except: pass return processes ``` The process information is included in the general report unless brief mode is selected: ```python info = { 'hostname': platform.node(), 'os': platform.system(), 'os_version': platform.version(), 'architecture': platform.machine(), 'python_version': platform.python_version(), 'uptime': get_uptime(), 'cpu': get_cpu_info(), 'memory': get_memory_info(), 'disk': get_disk_info(), 'network': get_network_info(), 'top_processes': get_process_info() if not brief else [] } ``` A normal invocation without section flags enables the complete report: ```python # If no specific option, show all show_all = args.all or not any([args.cpu, args.memory, args.disk, args.network, args.processes]) if args.json: print_system_info(json_output=True, brief=args.brief) elif show_all: print_system_info(brief=args.brief) ``` ### Technical Analysis The tool invokes `ps aux --sort=-pcpu` and extracts the owner, PID, CPU usage, memory usage, and full command line for the ten highest ...[truncated 2497 chars]
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding

The skill advertises command-line usage that implies shell execution and system inspection capabilities, but it does not declare any tool scope such as permissions or allowed-tools. That mismatch weakens security boundaries because an agent may invoke shell or file-reading behavior without an explicit least-privilege contract, increasing the chance of unintended host information disclosure.

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/sysinfo.py (reported line 102)May include surrounding context.

python
try:
        # Use df command
        result = subprocess.run(
            ['df', '-B1', '-T'], capture_output=True, text=True, check=True
        )
        lines = result.stdout.splitlines()[1:]  # Skip header

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill collects and prints per-process details including owning user and full command line, which can expose sensitive operational context far beyond basic system profiling. Command lines often contain secrets such as tokens, file paths, internal hostnames, or customer data, so this creates an information disclosure risk if the output is shared, logged, or shown to a lower-privileged user.

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/sysinfo.py (reported line 169)May include surrounding context.

python
processes = []
    
    try:
        result = subprocess.run(
            ['ps', 'aux', '--sort=-pcpu'], capture_output=True, text=True
        )
        lines = result.stdout.splitlines()[1:11]  # Top 10

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The code exposes sensitive process information without any user-facing warning, consent, or indication that command lines may contain secrets. While this is primarily a UX/privacy control gap rather than an exploit primitive, it increases the chance of accidental disclosure during diagnostics or automated logging.

Content

No source excerpt is available for this finding.