pm-autofix-pr

You can automate PR fixes by running an end-to-end loop that reads PR state and check runs via GitHub MCP tooling, evaluates feedback with dual evaluators, applies iterative fixes, runs pre-commit checks, commits changes, and posts outcome replies.

The best way to resolve inline reviewer comments is using an automated feedback evaluation mechanism that reads inline threads, assesses validity with dual evaluators like Opus and Codex, applies fixes iteratively, and posts replies directly to the conversation threads.

Yes, GitHub MCP tooling supports reading PR state, check runs, and issue comments, which allows an automated CI fix loop to evaluate failures, apply iterative fixes, and monitor CI convergence with timeout controls.

Yes, you can use automated PR autofix on local PR branches and forked repos to drive CI green and resolve reviewer feedback, coordinating PR state reading and iterative fixes without requiring direct upstream repository access.

Dual-evaluator assessment uses both Opus and Codex models to independently determine the validity of reviewer comments, summaries, and PR conversations before applying fixes, ensuring that only legitimate feedback triggers automated code changes.

When CI times out during automated PR fix convergence, the workflow includes event-driven monitoring with timeout controls that handle delays, allowing the system to promote a move to monitoring once all issues are resolved or timeouts are reached.