The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill performs local checks to identify if sensitive files (e.g., .env, .pem, .key, .credentials) are staged for commitment. This is a defensive security practice. It also references a local configuration file ~/.codex/config.toml for commit attribution settings.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute standard git and gh (GitHub CLI) commands. These operations are limited to the scope of repository management as described in the skill's purpose.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill ingests untrusted data from the local repository environment via git status, git diff, and git log commands.
Boundary markers: No explicit delimiters are used to separate ingested data from agent instructions.
Capability inventory: The skill has the ability to execute shell commands and create pull requests on remote repositories via the gh tool.
Sanitization: No explicit sanitization of git output is performed before it is presented to the agent, though the use of shell heredocs (cat <<'EOF') in the provided examples helps prevent unintended shell expansion of the content.

commit-push-pr