grading-frontend
Audited by Socket on Nov 12, 2026
2 alerts found:
Securityx2This script performs a Lighthouse performance audit and writes its results, but unvalidated TEAM input enables output path traversal and JavaScript injection in the node -e invocation, potentially allowing local code execution under the runner’s account. The caller-controlled URL also permits browser requests to arbitrary destinations, including internal services, and npx may download and execute Lighthouse. Severity is elevated when arguments are attacker-controlled; no specific external domain, credential target, or exfiltration endpoint is hardcoded.
This accessibility helper runs @axe-core/cli against a caller-supplied URL and writes report and summary files. A crafted TEAM argument can cause path traversal and JavaScript injection through the unescaped REPORT interpolation in node -e, potentially executing code with the script user's privileges. npx also executes the CLI package. Severity is high when arguments are attacker-controlled; no hardcoded external destination or credential theft is shown.