The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process multiple external data sources which may contain untrusted content.
Ingestion points: The workflow reads from HIERARCHICAL_PRODUCTION_PLANNING.md, VISIBLE_DIALOGUE_SYNC.md, SHOOTING_SCRIPT.md, screenplays, line manifests, blocking documents, and shot briefs.
Boundary markers: The instructions do not define clear delimiters or "ignore embedded instructions" warnings for the data being parsed.
Capability inventory: The skill writes results to the file system (03_preproduction/shooting_script.json) and executes the production-coverage command-line tool.
Sanitization: No sanitization or validation logic is specified for the text content extracted from input files.
[COMMAND_EXECUTION]: The workflow concludes by executing the production-coverage tool. While this is a standard step for the skill's filmmaking purpose, it constitutes a capability that could be misused if coupled with a successful injection attack.

shooting-script