Security audit
Security checks for vulnerabilities and agentic risk
The scanners classify this skill as benign because its flagged behaviors appear related to its purpose, while noting installation and privacy risks; confidence is medium and the scan is partial.
Consider whether prompts and results might contain private or proprietary information before sharing them, and review the exact material before publication. The installation example does not pin the package version. The scan inspected all five listed components but is marked partial because two references could not be resolved, so the evidence is incomplete.
The installation example invokes an unpinned npm package through npx, allowing the resolved package version to change and execute code with the installing user's permissions. This is a conditional supply-chain risk rather than evidence of malicious behavior; the finding incorrectly describes the command as an MCP server reference.
The workflow recommends publishing the complete prompt and full results without checking for confidential material or obtaining approval for those specific disclosures. Public comparison posts are appropriate to this skill, and no automatic upload is shown, but following this recommendation with private prompts or outputs could expose sensitive information unnecessarily.
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
def run(cmd):
r = subprocess.run(cmd, capture_output=True, text=True)
if r.returncode:
sys.exit(f"failed: {' '.join(cmd[:6])}…\n{r.stderr[-1500:]}")
return r.stdout