The Agent Skills Directory

[SAFE]: The provided Python scripts (design_critique.py, journey_mapper.py, usability_scorer.py) exclusively use the Python standard library. They do not perform network requests, access sensitive files, or execute system commands.
[INDIRECT_PROMPT_INJECTION]: The skill defines an ingestion surface for external data via JSON and CSV files.
Ingestion points: Data is loaded from file paths provided to command-line arguments (e.g., --answers, --stages, --sus-responses).
Boundary markers: None present in the parsing logic.
Capability inventory: The tools are restricted to data analysis and text formatting for report generation. No dangerous capabilities such as subprocess calls, network exfiltration, or arbitrary file system writes were identified.
Sanitization: Input is processed through standard json and csv library parsers.
[COMMAND_EXECUTION]: While SKILL.md documents several CLI commands, these are calls to the skill's own local Python scripts for legitimate design tasks (e.g., generating CSS from tokens or running accessibility checks). These do not represent a command injection risk as they are static examples of intended tool usage.

product-designer