post-edit-verification

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using python -c to validate JSON structure and verify field invariants, such as ensuring the description field is non-empty and under 100 characters. It also utilizes grep for targeted verification of specific file segments.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes component files that could contain untrusted data while maintaining capabilities to execute shell commands.
  • Ingestion points: Component JSON files loaded via file paths provided to validation commands.
  • Boundary markers: None specified; the skill performs direct parsing of the file content.
  • Capability inventory: Execution of Python code and the grep utility; file modification via standard editing tools.
  • Sanitization: The skill relies on standard JSON parsing logic to handle file contents.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 01:11 PM
Security Audit — agent-trust-hub — post-edit-verification