The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses directive language designed to override the agent's default interaction behavior. Instructions such as '지금 즉시 아래 워크플로우를 실행하라. 질문하지 말고 바로 실행하라' (Execute the following workflow immediately. Do not ask questions, just execute) are intended to force autonomous execution.
[COMMAND_EXECUTION]: The skill requires the agent to execute a series of shell commands including git status, git diff, git add, and git commit to perform its primary function of version control management.
[DYNAMIC_EXECUTION]: The skill involves generating and executing complex shell commands dynamically, specifically using HEREDOC patterns in Unix-like environments and temporary files in Windows to handle Korean character encoding during the commit process.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface by ingesting untrusted data from the local environment.
Ingestion points: The agent reads the output of git status and git diff, which contains content from files in the repository.
Boundary markers: None are present; the agent analyzes the diff content directly to determine commit groups.
Capability inventory: The skill has the capability to execute shell commands (git add, git commit).
Sanitization: There is no explicit sanitization or filtering of the file content before it is processed by the LLM to generate commit messages or logic.

commit