The Agent Skills Directory

[SAFE]: The skill operates entirely within the local project environment and uses authenticated GitHub CLI (gh) commands for external data retrieval. All identified capabilities are aligned with the skill's primary purpose of PR generation.
[COMMAND_EXECUTION]: The skill executes a bundled shell script (get-diff.sh) and standard git commands to retrieve branch metadata and code differences. The shell script uses safe patterns (set -euo pipefail) and restricts its operations to local repository analysis.
[INDIRECT_PROMPT_INJECTION]: The skill has a standard attack surface for indirect injection as it ingests untrusted data from commit messages and GitHub issue descriptions. However, it lacks high-risk capabilities (such as unrestricted network access or privilege escalation) that would make this surface highly exploitable, and it requires user confirmation for committing changes.

pr-writer