The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes logic from a local file scripts/editor_methods.py to perform consistency checks and report generation. This is standard behavior for an orchestration skill within a controlled workspace.
[DATA_EXPOSURE]: The skill interacts with a local database to query and update book sections and reads from local files like BRD.md. These operations are limited to the workspace and aligned with the skill's purpose of managing a writing project.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-generated draft content from the database. While this represents a potential injection surface if the content contains adversarial instructions, it is a standard risk for document processing agents and is handled within the context of the user's own workspace data.

bookstrap-edit-path