The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands via the scripts/smart_commit.sh script to perform git operations, including staging, committing, and pushing changes.
[DATA_EXFILTRATION]: The script uses the command git add . to stage all changes in the current directory. This creates a risk of accidental data exposure if sensitive files, such as environment variables (.env), cloud credentials (.aws/), or private keys, are present in the workspace and not explicitly excluded via a .gitignore file.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided commit messages which are passed directly to the shell script, creating an injection surface.
Ingestion points: The commit message is ingested via a command-line argument in scripts/smart_commit.sh (e.g., MESSAGE="${1:-chore: update code}").
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the script or instructions.
Capability inventory: The script has the capability to modify the repository state (git add, git commit) and communicate with remote servers (git push).
Sanitization: The commit message is not sanitized or validated before being used in the git commit command.

git-pushing