principle-sequence-verifiable-units

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The 'Uni-pstack Runtime Adapter' section contains explicit instructions to override the agent's default behaviors, model choices, and panel settings.
  • Evidence: 'Apply these overrides before following the original workflow'
  • Evidence: 'Replace upstream Composer, Claude Opus, and other panel defaults with Codex gpt-5.6-sol'
  • Evidence: 'Do not infer Claude model choices from this skill'
  • [COMMAND_EXECUTION]: The instructions direct the agent to execute external scripts and CLI tools directly via the host terminal, which bypasses standard tool constraints and relies on potentially unverified local scripts.
  • Evidence: 'launch Codex CLI workers using the installed pstack skill spawn-codex-worker script or direct codex exec'
  • Evidence: 'Use the host terminal, browser, review, subagent, and git tools directly'
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 04:46 PM