The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands to manage Git workflows, including git checkout, git merge, git rebase, and git commit. It specifically uses git rm -rfq <DIR> to delete directories identified as build artifacts (e.g., dist/) during high-confidence automated resolution.
[COMMAND_EXECUTION]: The skill uses the --no-verify flag during git commit. This bypasses pre-commit hooks which may be configured in the local environment to perform security scanning, linting, or secret detection.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8). It ingests untrusted data by reading raw file contents using git show and git diff for 'Semantic Analysis'. Maliciously crafted content within the files being merged (such as comments or strings) could attempt to influence the agent's resolution logic, confidence scoring, or instruction adherence during the analysis phase.
Ingestion points: Reads code from files using git show :1/2/3:<FILE> and git diff in the Y.2 Semantic Analysis phase.
Boundary markers: None. The raw content of the files is processed directly by the AI to determine 'intent'.
Capability inventory: The skill has the capability to delete files (git rm), commit changes (git commit), and manipulate the repository state.
Sanitization: None. The skill relies on the AI to distinguish between code and instructions within the processed files.

git-conflict-resolve