The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing local command-line tools such as git, pylint, flake8, eslint, black, and prettier via the subprocess.run method in several Python scripts (branch_manager.py, pre_commit_check.py, generate_commit_message.py, change_analyzer.py). These operations are consistent with the skill's intended purpose as a developer productivity tool.
[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection (Category 8) because it processes untrusted data from the local repository and passes it to the agent without sufficient boundaries.
Ingestion points: The agent ingests external data through git diff output in scripts/generate_commit_message.py and reads raw file contents to check for merge conflicts in scripts/pre_commit_check.py.
Boundary markers: The instructions in SKILL.md lack explicit boundary markers or directives to ignore instructions that might be embedded within the code or diffs being analyzed.
Capability inventory: The skill has the capability to execute a variety of git commands, including creating, switching, and deleting branches, as well as running external linting and formatting binaries.
Sanitization: There is no evidence of sanitization or filtering to prevent malicious instructions hidden in code comments or documentation from influencing the agent's behavior during the commit generation or analysis phase.

git-workflow-management