The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from git conflict stages, which could potentially contain malicious instructions. \n
Ingestion points: scripts/classify-conflicts.sh and scripts/take-side.sh read unmerged file versions from the git index. \n
Boundary markers: The skill relies on standard git conflict markers and diff output formatting to delimit content. \n
Capability inventory: The skill can write files to the repository and stage them via git add. \n
Sanitization: No explicit sanitization of the file content is performed. \n- [EXTERNAL_DOWNLOADS]: The Python script scripts/conflict_shape.py specifies a requirement for Python 3.14. The uv tool may download this version from official Python distribution sources if it is not found on the local system. \n- [COMMAND_EXECUTION]: The skill uses shell scripts to automate git operations and includes a PreToolUse hook (scripts/guard-resolve.sh) to block dangerous git flags and prevent staging files with unresolved markers. \n- [DYNAMIC_EXECUTION]: An internal verification step in scripts/resolve-union.sh executes a hardcoded Python script via a heredoc to verify the integrity of the resolved content.

resolve-rebase-conflicts