The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various Git and GitHub CLI commands to manipulate the local and remote repository state. Specifically, it modifies the repository's configuration (e.g., setting merge.conflictStyle to zdiff3 and enabling rerere) and is authorized to push changes to remote branches and comment on pull requests.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it involves reading and processing untrusted data from files that are currently in a conflicted state.
Ingestion points: File content is ingested via the Read tool for all files returned by git diff --name-only --diff-filter=U.
Boundary markers: There are no explicit boundary markers or system-level instructions to ignore potential commands embedded within the code or text being resolved.
Capability inventory: The skill possesses significant capabilities, including the ability to Edit files, git commit, git push changes to remote repositories, and gh pr comment on GitHub.
Sanitization: No content sanitization or validation is performed on the data read from the files. While the instructions require the removal of conflict markers, they do not provide safeguards against malicious instructions embedded within the sections of the code being merged.

git-conflicts