The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute several local Python scripts, including prepare_monolith_input.py, reassemble_chunks.py, restore_modality.py, strip_injected_commas.py, and verify_gates.py. These scripts handle core functionality such as text preprocessing, chunk reassembly, and quality assurance gating.
[DYNAMIC_EXECUTION]: The skill dynamically calculates its installation root directory (SKILL_ROOT) by traversing the parent directories of the platform-provided ${CLAUDE_SKILL_DIR} until it identifies a specific marker file. This calculated path is then used as a base to invoke internal scripts, ensuring portability across different deployment environments.
[INDIRECT_PROMPT_INJECTION]: As the skill is designed to process and rewrite user-provided Korean text, it is exposed to potential indirect prompt injection attacks. The instructions explicitly mitigate this risk by mandating that input text be treated strictly as data and that any imperative commands or overrides embedded within the user text must be ignored.

fluent-korean