eval() with non-literal argument

What it detects

eval() with non-literal argument: an injection check run by the textscan analyzer.

Evidence

The matching source path, line, and bounded code observation produced by the native rule.

Remediation

Replace eval() with explicit parsing (JSON.parse for JSON, Function constructor with allowlist for templated logic). Eval of user-controlled input is direct RCE.

Limitations

This check is limited to the applicability and evidence delivered by textscan. Review the evidence in context; a match is a finding input, while confidence and policy determine its disposition.

Related reference

Analyzer Registry · Findings and release decisions · Policy and confidence

Remediation guidance: Remediation and fix verification · Baselines and suppressions