The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted HTML data provided by users (files or pastes) to recreate them within WordPress builders. This creates an attack surface where malicious instructions hidden in the HTML source (e.g., acrostics, hidden comments, or deceptive text) could influence the AI agent's actions during the conversion or validation steps.
Ingestion points: SKILL.md identifies user-supplied HTML files and pastes as the primary input for the respira_convert_html_to_builder tool.
Boundary markers: The instructions do not specify the use of clear delimiters or instructions to ignore embedded prompts when handling the HTML input.
Capability inventory: The skill utilizes tools to create/modify WordPress pages, upload media to the server, and generate share links for visual preview.
Sanitization: The instructions explicitly note that 'Scripts are not carried,' which acts as a sanitization layer against JavaScript-based attacks in the rendered output, although it does not inherently prevent the AI from obeying Natural Language instructions embedded in the code.
[EXTERNAL_DOWNLOADS]: The conversion workflow fetches linked stylesheets and font resources from external URLs via HTTPS to maintain the visual fidelity of the resulting page. This is a functional requirement for the migration process.

html-to-page