The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several local Python scripts (e.g., bootstrap_progress_screen.py, render_progress_screen.py, check_progress_consistency.py) to manage and visualize task states. These scripts are invoked via the command line using environment variables such as $UIUX_SKILL_SYSTEM_ROOT and $PROJECT_ROOT, which is standard for integrated developer tools.
[SAFE]: The skill includes robust safety guidelines in references/safety-boundaries.md that explicitly forbid the inclusion of sensitive data such as API keys, tokens, recovery codes, or environment variable values in the progress dashboard. It also implements a mandatory no_secret_scan.py check to enforce this policy.
[SAFE]: Data handling is restricted to a 'local-first' approach. The instructions specify that the progress screen is a derived view and must not be published without explicit approval and sanitization, reducing the risk of accidental data exposure.
[PROMPT_INJECTION]: The skill processes untrusted input from project-request.json and interpolates user-provided data into script arguments. While this creates a potential surface for indirect prompt injection, the risk is mitigated by the skill's focus on local visualization and its integrated validation/scanning workflows.

agent-progress-visualizer