The Agent Skills Directory

[SAFE]: The skill provides documentation and examples for standard web development practices related to HTML validation.
[EXTERNAL_DOWNLOADS]: The skill suggests installing the html-validate package via npm and utilizing official GitHub actions (such as actions/checkout and actions/setup-node). These are standard, well-known development dependencies and workflows.
[DATA_EXFILTRATION]: The instructions include examples of sending HTML markup to the official W3C validator service (validator.w3.org) via curl or integrated browser tools. This is the primary intended functionality of the skill for validation purposes and does not target sensitive files like credentials, SSH keys, or environment variables.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process HTML markup, which is an external data source. While this represents a potential attack surface for indirect prompt injection, the validation tools recommended (such as html-validate and the W3C Nu Validator) are static analysis tools that do not execute the processed content. The skill's own instructions focus on reporting and fixing structural errors, posing a low risk to the agent's integrity.

w3c-compliant