ui-implementation
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local validation tools, specifically
node scripts/ui_governance_guard.mjsandnode scripts/accessibility_governance_guard.mjs. These appear to be project-specific linting or testing scripts. - [INDIRECT_PROMPT_INJECTION]: The workflow involves reading multiple local files, including architectural decision records (ADRs), UI inventories, and pattern manifests. These files serve as the ground truth for UI implementation and governance within the project.
- [SAFE]: The skill manages state and configuration via local environment variables (e.g.,
CLAWIX_UI_GUARD_DIFF_BASE) and does not perform any external network requests, data exfiltration, or obfuscation. The behavior is consistent with standard software engineering practices for a large-scale UI codebase.
Audit Metadata