The Agent Skills Directory

[SAFE]: The skill is a developer utility for the Agent Skills ecosystem, designed to assist in the creation and maintenance of other skills.
[COMMAND_EXECUTION]: Several Python scripts perform file system operations and execute local git commands.
scripts/bundle_skill.py uses subprocess.run to call git log, git show, and git ls-tree. These calls use list arguments, specify a working directory, and do not use shell=True, which is a safe implementation for repository-level operations.
scripts/init_skill.py creates directories and files using pathlib and sets permissions with chmod 0o755, which is standard for making scripts executable.
[DATA_EXPOSURE]: No sensitive data access or exfiltration patterns were detected. The scripts operate entirely on local skill directories and project metadata.
[DYNAMIC_EXECUTION]: The skill uses yaml.safe_load in scripts/skill_frontmatter.py to parse metadata, avoiding the risks associated with unsafe YAML loading. No instances of eval() or exec() were found.
[REMOTE_CODE_EXECUTION]: There are no network operations (such as curl or requests) that download or execute external code. All resource resolution is performed against local directories.

create-skill