ce-commit-push-pr
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository metadata and history, such as commit messages and existing Pull Request descriptions, to generate new PR titles and bodies.\n
- Ingestion points: The skill ingests untrusted content via
git logandgh pr viewfrom the repository being processed (documented inreferences/context.mdandreferences/pr-description-writing.md).\n - Boundary markers: It employs quoted heredoc sentinels (
'__CE_PR_BODY_END__') when writing the PR body to a temporary file, which effectively prevents shell variable expansion and command substitution (documented inreferences/apply-and-handoff.md).\n - Capability inventory: The skill uses
gitandghCLI tools for repository operations.\n - Sanitization: The skill validates branch names derived from external sources against a safe regex (
[A-Za-z0-9._/-]+) to prevent command injection and escapes PR titles before they are passed to the shell (documented inreferences/stack-submit.mdandreferences/apply-and-handoff.md).\n- [COMMAND_EXECUTION]: The skill makes extensive use of shell commands to perform repository management and interact with the GitHub CLI.\n - Evidence: Reference files contain numerous shell snippets for branching, committing, pushing, and Pull Request lifecycle management.
Audit Metadata