The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute Python logic via shell blocks (`python3
<<'PY'`). This is the intended mechanism for the agent to interact with the skill's bundled orchestration scripts.
[EXTERNAL_DOWNLOADS]: The runner.py script performs a network availability check by attempting to open a /health endpoint on the service URL defined in the MKIVIDEOS_URL environment variable. This is used to decide between local and remote rendering pipelines.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core functionality of processing user-supplied narrative text.
Ingestion points: User-defined subjects, premises, and episode descriptions are stored in biblia.json and episode-specific JSON files (e.g., ep01.json).
Boundary markers: The skill does not employ specific boundary markers or delimiters when interpolating user-supplied content into downstream prompts for image and video generation tools.
Capability inventory: The skill possesses the ability to read and write to the local file system (assets and manifestos), execute Python code through the shell, and make network requests to local or configured media services.
Sanitization: There is no evidence of robust sanitization or filtering of user-provided strings before they are incorporated into prompts for generative AI services (e.g., inemaimg or inemavox).

inemaref-serie