The Agent Skills Directory

[SAFE]: The skill is entirely informational, providing technical reference data and methodology for architectural design. No security vulnerabilities were found.
[PROMPT_INJECTION]: The instructions are strictly limited to providing material selection guidance. No patterns of jailbreaking, system prompt extraction, or instruction overrides are present.
[DATA_EXFILTRATION]: There are no commands for file system access, credential harvesting, or network communication. The skill does not handle or transfer sensitive data.
[REMOTE_CODE_EXECUTION]: The skill does not reference or download external packages, scripts, or binaries. There is no infrastructure for executing remote code.
[NO_CODE]: All analyzed files consist of pure Markdown documentation and reference tables. No scripts (Python, JavaScript, shell) or executable logic are included in the skill package.

material-selection