The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local scripts and Git commands to manage the commit workflow. It specifically invokes node core/hooks/_shared/receipt-mint.js to record review verdicts and performs multiple sequential commits. These operations target the local repository environment using standard development tools.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it ingests and processes the content of staged and unstaged files from the working tree to generate commit messages and grouping logic.
Ingestion points: The skill reads staged and unstaged files from the full working tree as described in the SKILL.md description.
Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish between file content and its own processing instructions, which could allow maliciously crafted code comments to influence the commit plan.
Capability inventory: The skill is capable of executing shell commands through node and performing repository modifications via Git.
Sanitization: The instructions do not specify any sanitization, filtering, or validation of the file content before it is interpolated into the agent's context for analysis.

smart-commit