implementation-runner

To automate in-repo code changes and generate a patch for review, use an implementation runner to execute planned steps, track changed files, and output a patch list. It records the final completion status and notes directly in the analysisContext for traceability.

Logging implementation outcomes in analysisContext works by automatically summarizing key code changes and recording the final completion status into the context object. This provides clear failure handling and prevents unintended interactions by documenting results centrally for auditability.

Yes, you can use an implementation runner in a chain-based workflow to track changed files. It is specifically designed for chain workflows where a user request requires applying changes, automatically recording the changed files and generating a patch for review.

No, you do not need external automation tools to execute real implementations within your repository. The implementation runner enforces no external skill invocations, coordinating actual implementation steps internally and recording the completion status directly into the analysisContext.

Limitations of using an implementation runner for code changes include its enforcement of no external skill invocations, restricting interactions to the prepared analysis context. It strictly handles in-repo implementations and outputs a patch list, requiring a pre-existing chain workflow to function properly.