The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user prompts to generate tailored explanations and orientation content. While this creates a standard interaction surface for instruction-based responses, the skill's logic is focused on output formatting (STE style, diagrams, HTML). As it interprets user queries like 'explícame Docker' to generate content, standard platform guardrails for LLM interactions apply.
Ingestion points: User prompts requesting explanations, register changes, or project status (e.g., 'no entiendo', 'explícamelo', 'háblame más técnico').
Boundary markers: The skill relies on the agent's internal instruction following; no specific data delimiters are defined in the instructions.
Capability inventory: File system access to read and update the user profile (02-DOCS/wiki/harness/user-profile.md) and the ability to generate and open local HTML files (open path/to/explain-<topic>.html).
Sanitization: Not explicitly defined in the skill logic.
[EXTERNAL_DOWNLOADS]: The documentation and test cases mention the use of external resources and tools, including the ElevenLabs API for video narration and the author-owned @ericrisco/rsc package. These are recognized as well-known services and vendor-specific resources respectively, presenting no inherent security risk to the agent environment.
[COMMAND_EXECUTION]: The skill facilitates the creation and opening of HTML-based explanation pages using the open command. This is used for orientation purposes and operates within the local file system scope defined by the user's project.

orient