The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes instructions that explicitly override external design guidance (specifically typeface selection rules from other potential skills) to ensure adherence to Material Design 3 standards (SKILL.md).
[PROMPT_INJECTION]: Indirect Prompt Injection Surface. The audit command processes data from untrusted external websites and local source code files.
Ingestion points: External URLs visited via browser tools and local file paths provided to the audit command (SKILL.md).
Boundary markers: Absent. The instructions do not define delimiters or warnings to treat ingested content as data rather than instructions.
Capability inventory: Browser navigation, DOM reading, JavaScript execution in browser context, file reading, and shell command execution (SKILL.md).
Sanitization: Absent. No verification or sanitization of ingested content is specified.
[COMMAND_EXECUTION]: The skill provides shell commands using grep for the agent to scan local project files for non-compliant design patterns during audits (SKILL.md).
[EXTERNAL_DOWNLOADS]: The skill references and guides the installation of official Google packages (@material/web and @material/material-color-utilities) from the npm registry and utilizes official documentation from m3.material.io.

material-3