The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for linting and outlining purposes.
Evidence: SKILL.md contains instructions to use uv run to execute Python scripts (lint.py, outline.py) against generated files.
[DYNAMIC_EXECUTION]: The skill uses a dynamically defined environment variable to locate and execute external scripts.
Evidence: SKILL.md defines CORE_SKILL_DIR as a sibling directory (natural-korean) and uses it to construct the path for script execution.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted meeting transcripts and memos, which could contain malicious instructions.
Ingestion points: The skill is designed to process external inputs such as "녹취·메모" (recordings/memos) and "인터뷰·워크숍 기록" (interview/workshop records) as defined in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions that might be embedded within the processed meeting content.
Capability inventory: The skill provides the agent with the ability to execute shell commands via uv run and write to the file system using provided templates in assets/.
Sanitization: No evidence of input sanitization or filtering of the processed text is found in the instructions.

write-korean-meetings-slides