resolve_pr_parallel

To resolve GitHub pull request review comments in parallel, this skill dispatches a dedicated background agent for each unresolved thread, implements the necessary code changes concurrently, commits them, and programmatically resolves the threads via the GitHub API.

Parallel processing of PR review feedback works by spawning individual agents for each unresolved comment thread, allowing simultaneous code changes across different files, which significantly speeds up the code review feedback loop.

Yes, you can use this to fix code review comments across multiple files simultaneously. It assigns each comment to a separate agent, processes them concurrently, commits the changes, and resolves the threads automatically.

You do not strictly need to specify a PR number to resolve review threads automatically. The skill can detect the current Git branch and its associated pull request, but it also supports working with a specified PR number if provided.

After pull request comments are resolved, the agents write detailed resolution reports to disk. This disk-based output prevents context exhaustion and allows you to review the automated changes and progress later.

Resolving PR comments in parallel prevents context exhaustion because the dispatched agents persist their resolution reports to disk rather than keeping all thread data and code changes in active memory.