The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external reference documents and manifests to generate project files, which presents a surface for indirect prompt injection.
Ingestion points: Reads from ../../references/EDITOR_PROJECT_EXPORT.md and uses manifest files provided at runtime.
Boundary markers: The instructions lack delimiters or warnings to the agent regarding potential instructions embedded within the editorial data.
Capability inventory: The skill calls a local Python script which performs file-system operations.
Sanitization: No sanitization or validation of the input project data is defined in the workflow.
[COMMAND_EXECUTION]: The skill requires the execution of a Python script (scripts/editor_project_export.py) to perform its primary function.

shotcut-export