The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to run npx --yes notugly, which automatically downloads a package from the public npm registry at runtime without requiring user confirmation.
[COMMAND_EXECUTION]: The skill relies on shell command execution to perform its primary function of contrast ratio calculation and vision simulation via the notugly CLI.
[REMOTE_CODE_EXECUTION]: By using npx, the skill downloads and executes code from a remote repository (npmjs.org) that is not part of a pre-defined trusted vendor list, effectively performing remote code execution during the auditing process.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data that is directly interpolated into shell commands, creating a surface for command injection attacks.
Ingestion points: User-provided URLs for the onepager command and hex code values for the fix command in SKILL.md.
Boundary markers: No delimiters, markers, or warnings are present to isolate the untrusted user data from the command string.
Capability inventory: The skill possesses the capability to execute shell commands via the npx utility as defined in the "Programmatic Helper" section.
Sanitization: The instructions do not specify any validation, escaping, or sanitization logic for the input parameters before they are passed to the shell environment.

accessibility-audit