ui-percept-rapid-feedback

Pass

Audited by Gen Agent Trust Hub on Mar 22, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell scripts (run-feedback.sh, archive-round.sh) to orchestrate UI testing workflows, including invoking xcodebuild for captures and uv run for tool execution.
  • [COMMAND_EXECUTION]: A Python script is embedded within run-feedback.sh to parse JSON analysis data and generate markdown reports and decision logs on the local filesystem.
  • [EXTERNAL_DOWNLOADS]: The archive-round.sh script supports an optional --push flag which executes git push to synchronize local UI capture history with a remote repository.
  • [PROMPT_INJECTION]: The skill generates specific prompts for the agent to assume a persona (iOS UX specialist) and provides a surface for indirect prompt injection by interpolating tool output into these prompts.
  • Ingestion points: lint.json, signals.json (generated by ui-percept).
  • Boundary markers: None.
  • Capability inventory: Shell and Python execution, Git operations, and instructions for code patching.
  • Sanitization: None; data is interpolated directly from JSON.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 22, 2026, 07:18 PM