Security checks across malware telemetry and agentic risk
This skill is not clearly malicious, but it gives a conversational AGI companion broad local file, process, system-info, and shell-command powers without tight safeguards.
Treat this as a Review install. Use it only in a sandbox or disposable workspace, keep secrets and important files out of reach, and do not enable the command/file/process tools unless you intentionally want the skill to have host-administration-level power.
start_time = datetime.datetime.now()
try:
if detected_shell_type == "powershell":
result = subprocess.run(
[shell_executable, "-Command", command],
cwd=cwd,
env=env,timeout=timeout
)
else:
result = subprocess.run(
command,
shell=True,
cwd=cwd,elif tool_name == "calculator":
expression = params.get("expression", "0")
try:
result = eval(expression)
data = {
"expression": expression,
"result": result67/67 vendors flagged this skill as clean.