The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructs the agent to fetch content from an external source to guide its output structure, creating a surface for indirect prompt injection.
Ingestion points: External gallery and examples located at thariqs.github.io/html-effectiveness/ and its subdirectories.
Boundary markers: Absent. The skill directs the agent to "shape the output after that example's structure and interactivity" without isolating the fetched data or using delimiters to ignore embedded instructions.
Capability inventory: The skill performs file system writes to the html-expl/ directory and executes the open shell command.
Sanitization: Absent. The skill does not provide instructions to validate or sanitize the external HTML content before processing it.
[EXTERNAL_DOWNLOADS]: The skill fetches structural templates and reference examples from thariqs.github.io. Additionally, it references assets from well-known services, specifically cdn.jsdelivr.net for the Mermaid library and fonts.googleapis.com for typography.
[COMMAND_EXECUTION]: The skill uses the open command to launch the system's default browser on newly created HTML files. While this is a standard operational step for the skill's purpose, it involves executing a shell command on files created by the agent.

explain-via-html