The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard git commands (git diff, git log, git show) to retrieve repository metadata and file changes. These commands are used in a read-only manner and are essential for the skill's primary functionality of generating commit messages.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository by reading file diffs. This constitutes a potential attack surface if malicious instructions are embedded in the code changes being analyzed, though the risk is inherent to tools that analyze user-provided code.
Ingestion points: Workflow steps in SKILL.md that read output from git diff --staged and git log.
Boundary markers: Absent; the diff content is directly incorporated into the agent's context without specific delimiters or isolation instructions.
Capability inventory: The skill is restricted to read-only git commands and lacks network access or file system modification permissions.
Sanitization: The skill does not perform validation or sanitization on the content of the diffs.

conventional-commits