cfe-validate

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes a PowerShell script (cfe-validate.ps1) to check the structural integrity of 1C extensions. The script includes an optional OutFile parameter that allows writing the validation results to a user-specified path, which could potentially be used to overwrite sensitive system or configuration files.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it extracts and displays data from user-provided XML files. An attacker could place malicious instructions within the XML metadata that the agent might follow upon seeing the validation output.
  • Ingestion points: Configuration.xml and other associated XML files within the directory specified by ExtensionPath.
  • Boundary markers: None present. The report merges validated data names and property values directly into the output stream.
  • Capability inventory: Subprocess execution (PowerShell), file system read (XML parsing), and file system write (via OutFile).
  • Sanitization: None. The script does not validate or sanitize strings extracted from the XML files before including them in the final report.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 08:36 AM