PR Draft Summary

PR Draft Summary is an official skill from the openai/openai-agents-python repository. It defines a controlled closing workflow for development tasks where a pull request draft is needed. The primary source is the official OpenAI Agents SDK path .agents/skills/pr-draft-summary. According to the provider, the workflow first collects the current branch, working tree status, untracked files, changed paths, size information, the latest release tag, and the appropriate base commit. It then classifies the change as a feature, fix, refactor, or behavior-relevant documentation change and produces a copy-ready block.

Purpose and output

The result is not an automatic publication and not a commit. According to the provider, the skill produces a concise summary, a branch-name suggestion, a single PR title, and a description that starts with wording such as “This pull request adds”. The draft should explain what changed, why the change was needed, which behavior considerations matter, and whether a backward-compatibility risk exists. When no changes or commits are found relative to the base, the workflow should not manufacture a pull request block. That boundary prevents an empty working state from being presented as completed implementation work.

Repository workflow

The skill is designed for repository-aware tasks and requires checking the actual Git state. Branch, upstream, and base reference are resolved from the local repository. The comparison considers both working-tree changes and commits ahead of the base. Untracked files are checked explicitly because a diff summary alone can omit them. Provider-defined category signals include runtime code, tests, examples, build and test configuration, and documentation with behavioral impact. The resulting draft should emphasize only the most important changed paths, keeping it easy for a human contributor to inspect.

Safety, limits, and accountability

PR Draft Summary writes text and suggests a branch, but it does not create a branch, commit, push, or open a pull request. This separation matters because a draft is not authorization for an external or durable action. Secrets, tokens, private diff contents, and confidential comments should not be placed in stored catalog copy. Actual data transmission depends on the selected agent, model route, and enabled tools; a local agent does not automatically mean that model processing remains local. The skill can summarize Git metadata, but subject-matter accuracy, test quality, security approval, and review accountability remain with the team.

The instructions normalize issue and pull request references for GitHub paste readiness while avoiding unnecessary direct actions. They can give an explicitly requested PR draft priority, while skipping trivial conversation tasks, repository metadata-only work, and documented exclusions. The official OpenAI Agents documentation provides the broader SDK context for agentic applications; this entry’s concrete workflow comes from the named skill file. The repository publishes its source under the MIT license. This catalog entry was reviewed on September 9, 2026 against the official skill file, the official repository, and OpenAI Agents documentation. The skill does not replace human approval and does not store GitHub stars because the Skill model has no github_stars field. Before sharing a draft, teams should also confirm that branch names, issue references, internal paths, and release conventions match their own governance. The draft is a communication aid and must always be checked against the repository’s final state.