Security audit
Security checks for vulnerabilities and agentic risk
ClawScan rated this skill suspicious because the scanners flagged an unverified downloaded installer and environment-controlled command execution; the review is based on incomplete analysis.
Review the installer and its system changes before running it, and avoid installing if you cannot verify the installer through a trusted source. Consider whether the skill's undeclared tool scope and environment-based command overrides fit your setup. Skillspector inspected all listed components, but marked analysis partial because three local path-like references could not be resolved; the evidence is therefore incomplete.
Without declared permissions the skill's intent is opaque and cannot be validated.
The instructions immediately execute a remotely downloaded, mutable installer without verifying its integrity or authenticity beyond HTTPS. Compromise of the distribution endpoint or artifact could execute arbitrary code with the installing user's privileges, potentially exposing cloud credentials; the read-only API restrictions do not constrain this installation step.
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
cmd.extend(extra)
cmd.extend(argv)
try:
proc = subprocess.run(
cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
timeout=120, universal_newlines=True,
)
Production execution accepts an arbitrary executable through VPCEP_LIST_HCLOUD_BIN and unchecked leading arguments through VPCEP_LIST_HCLOUD_ARGS, bypassing the read-only operation guard at the process boundary. If a less-trusted launcher or environment configuration can influence these variables, it can execute code with the agent's privileges and inherited credentials; this is conditional on environment control, not shell metacharacter injection.
cmd.extend(extra)
cmd.extend(argv)
try:
proc = subprocess.run(
cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
timeout=120, universal_newlines=True,
)
The workflow specifies Chinese hints for empty results, permission failures, invalid IDs, and missing credentials, regardless of the user's language. No language opt-in or justification for requiring Chinese diagnostics is provided; using Huawei Cloud alone does not justify this output-language constraint.
These criteria explicitly require Chinese hints for permission failures, credential errors, and invalid resource IDs. The file provides neither user opt-in nor a documented locale-specific reason for fixing these messages to Chinese.
Dynamic getattr() with a non-literal attribute name can access arbitrary object attributes, potentially bypassing access controls.
response_key = "endpoint_services"
try:
response = call(request)
batch_objs = list(getattr(response, response_key) or [])
except Exception as exc:
low = str(exc).lower()
if any(p in low for p in _AUTH_PATTERNS):