gpd-compare-branches

To compare results across hypothesis branches in a GPD workflow, this Skill extracts and synthesizes data from STATE.md, SUMMARY.md, and VERIFICATION.md to build a structured side-by-side comparison. It highlights gaps and flags potential tensions to accelerate your decision-making.

The best way to decide which hypothesis branch to merge or prune is to generate a formal BRANCH-COMPARISON report. This report aggregates verification status and context usage across branches, computing numerical differences to guide your pruning or merging decisions.

Yes, you can automatically extract verification status and context usage from markdown files. The Skill aggregates key results from STATE.md, SUMMARY.md, and VERIFICATION.md across all branches to provide a holistic view of progress and alignment.

To identify tensions or disagreements across multiple research branches, the Skill computes numerical differences and flags potential conflicts during the cross-branch comparison. It extracts data from your verification files to highlight misalignments automatically.

Yes, you need specific markdown files to perform a branch comparison. The Skill requires STATE.md, SUMMARY.md, and VERIFICATION.md files to exist within your hypothesis branches to extract the necessary data for its structured cross-branch analysis.

You should export a formal branch comparison report when you need to document the decision to merge or prune hypothesis branches. The exported report captures the structured comparison, key results, and computed tensions for future reference.