The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to execute git commands and helper scripts (scripts/detect.py and scripts/lint_commit.py) to inspect the repository state, identify configurations, and perform the final commit operation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository, including commit history, branch names, and code diffs, to determine the appropriate commit message style. This content could potentially contain malicious instructions intended to manipulate the agent's logic.
Ingestion points: The script scripts/detect.py reads previous commit subjects and bodies via git log, and SKILL.md instructs the agent to read the current git diff.
Boundary markers: There are no explicit instructions or delimiters used when analyzing the repository content to ensure the agent ignores any instructions embedded within the analyzed data.
Capability inventory: The skill possesses the ability to execute shell commands (git commit, git add) and write files to the .git directory (COMMIT_BODY.md).
Sanitization: No sanitization or filtering is performed on the commit history or diff data before it is presented to the agent for analysis.

commit-changes