The Agent Skills Directory

[SAFE]: No security issues were detected. The skill is dedicated to providing UI accessibility guidelines (WCAG 2.2) and a utility for color contrast verification.
[COMMAND_EXECUTION]: The skill includes two Python scripts (contrast-check.py and contrast-mcp.py) that the agent is encouraged to run via the Bash tool. These scripts are transparent, use standard library functions, and include input validation (regex) to ensure that only valid hex color strings are processed, preventing command injection or other malformed input attacks.
[DATA_EXFILTRATION]: There are no network-capable commands (like curl or wget) or external URL references used for data transfer. All operations are local.
[PROMPT_INJECTION]: The instructions in SKILL.md are strictly limited to accessibility standards and interaction affordances. There are no attempts to bypass safety filters or override system-level instructions.
[INDIRECT_PROMPT_INJECTION]: While the skill processes UI color data, it implements strict input validation. The contrast-mcp.py tool uses re.fullmatch(r"[0-9A-Fa-f]{6}", h) to sanitize inputs before processing, which effectively mitigates the risk of processing malicious data through the tool.

antislop-human