json-content-management

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION] (MEDIUM): Unsafe dynamic import path construction in lib/data.ts. The loadData function interpolates the locale and filename variables directly into an import() call. Because the locale value is sourced from URL parameters in app/[locale]/services/page.tsx, an attacker could provide a malicious string (e.g., ..%2F..%2F) to attempt to load unauthorized files within the application's source directory.- [INDIRECT_PROMPT_INJECTION] (MEDIUM): Lack of input validation for data-driving parameters. This vulnerability allows untrusted external data to influence the skill's file-loading logic.
  • Ingestion points: The locale variable extracted from params in app/[locale]/services/page.tsx.
  • Boundary markers: Absent. There are no checks to ensure the locale matches a set of allowed values before being used in path construction.
  • Capability inventory: Uses import() which can load and parse JSON or execute JavaScript/TypeScript files included in the build context.
  • Sanitization: Absent. The code relies on TypeScript type assertions (as Locale) which have no effect at runtime and do not prevent malicious strings from reaching the file system operation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 07:52 AM