mhattingpete

Reads size, line count, modification dates, and content distribution for any file. Useful for spotting bloat, tracking changes over time, or understanding a codebase layout.

Takes a pattern (old function name, deprecated API, inconsistent variable style) and refactors it everywhere it appears in your codebase, with diffs you can review first.

Takes a repo and adds linting, testing scaffolds, CI/CD stubs, docs structure, and pre-commit hooks—all configured and ready to run, not just templated.

Reads your repo structure and produces a self-contained HTML site with API reference, code examples, architecture diagrams, and syntax highlighting—ready to share or host.

Generates an interactive HTML diagram that shows how data flows, where features live, what the stack is, and how it deploys — all in one view you can share or embed.

Takes a list of reviewer comments and produces a prioritized set of diffs that address each one, with explanations of what changed and why.

Runs your test suite, groups failures by root cause, and patches each class of error with the most likely fix — then reruns to confirm.

Reads your repo, maps the architecture and data flow, then writes a guide explaining what each layer does and how they talk to each other.

Stages, commits with a conventional message, and pushes to your remote in one go. Writes the commit message for you based on what changed.

Executes Python locally against multiple files or a complex workflow in one shot, then feeds results back to Claude. Cuts token spend 90% on tasks like batch transcoding, parsing, or refactoring.

Takes a written description of how something works and produces an interactive HTML flowchart with decision paths, color-coded stages, and labeled swimlanes you can edit inline.

Takes a list of milestones, phases, and dates and renders an interactive HTML timeline or Gantt chart with progress bars, phase groupings, and shareable embeds.

Generates 3–5 distinct approaches to a problem in parallel—different architectures, code styles, or creative directions—then scores and ranks them so you compare apples to apples.

Scans your repo for architecture debt, security gaps, test coverage blind spots, and performance drains. Returns a prioritized list of what to fix first and why.

Scans your Claude conversation logs and identifies repeated friction points, common mistakes, and moments where a skill or macro would save you time.

Converts a CSV or sheet of metrics into an interactive HTML dashboard with cards, charts, and gauges. No coding. Embeds or emails the result.

Takes a vague feature idea and breaks it into sequenced, implementable tasks with acceptance criteria, dependencies, and effort estimates.

Extracts a function, class, or snippet from one file and inserts it at an exact line in another. Handles imports and dependencies without duplication.