benchmark

To benchmark dotforge configuration quality, this Skill isolates full and minimal configurations in two separate git worktrees, runs standardized tasks, and compares execution results across stacks detected from .claude manifests. It surfaces configuration differences to guide governance decisions.

Comparing full versus minimal dotforge configurations requires a git repository containing .claude/settings.json, a CLAUDE.md file, and standardized prompts located under tests/benchmark-tasks. The Skill automates worktree creation and task execution once these inputs exist.

Git worktrees provide isolated environments to simultaneously test full and minimal dotforge configurations. By running identical benchmark tasks in each worktree, the Skill captures execution metrics that reveal how each configuration impacts reliability and maintenance.

Yes, this benchmarking approach supports multi-repo environments by detecting stacks from .claude manifests and applying standardized tests across repositories. It generates actionable insights to guide tooling decisions in environments with varying configuration depths.

Run a dotforge configuration benchmark when evaluating how full versus minimal configurations affect task execution and reliability. It is designed for teams making informed configuration and governance decisions across multiple stacks in a multi-repo environment.

Dotforge configurations behave differently due to varying stack requirements detected from .claude manifests. Benchmarking full versus minimal setups in isolated worktrees surfaces these execution differences, helping identify the most reliable configuration for each stack.