Developer Productivity

Turns your git commits into a customer-friendly changelog, grouped by features, improvements, security and fixes.

Checks whether work meets its acceptance criteria and stops there, without adding scope.

A guide for building high-quality MCP servers in TypeScript or Python: research, design, implementation, testing and evaluations.

Builds your project's shared vocabulary: challenges fuzzy terms, tests them with edge cases, and records them in GLOSSARY.md and ADRs.

A pull request body template: the smallest diagram that explains the change, before and after evidence, and a merge-danger call.

Control and inspect a live Chrome browser through the Chrome DevTools MCP: navigate, click, fill forms, take snapshots and screenshots, read the console and network, and profile performance.

Debug a failing GitHub Actions check on a pull request: find the failing checks, pull the logs, summarize the cause, draft a fix plan, and change code only after you approve.

Fixes a bug at the narrowest layer that owns it, with a regression test and no unrelated cleanup.

Create Playwright automation for inspecting, testing, and debugging local web applications.

A six-phase loop for hard bugs: build a reproduction that fails on exactly this bug, minimise it, test ranked hypotheses, then fix with a regression test.

Hand a finished piece of work to a fresh reviewer: a code-reviewer prompt with the plan and the exact git range, graded findings, and rules for acting on them.

Turn an idea into an approved design before any code is written: classify the task, ask one question at a time, compare approaches and get your sign-off at each stage.

Review a codebase like a security researcher: trace user input to dangerous sinks, find injection, access-control and secrets problems, and get patches to review, never applied automatically.

Makes your agent find and prove the cause of a bug before it changes any code.

Builds a throwaway prototype to answer one design question: a clickable HTML model for logic, or several UI variations to compare.

Compacts the current conversation into a handoff document so a fresh agent session can pick the work up.

Has your agent interview you in numbered rounds, with a recommended answer for each question, until every decision in your plan is settled.

Handle review feedback with technical rigour: understand it, check it against the code, push back when it is wrong, clarify before acting and implement one item at a time.

Write the failing test first, watch it fail, write the minimum code to pass, then refactor: the red-green-refactor loop with strict rules and a checklist.

A reference for writing documents agents read, such as skills, CLAUDE.md and AGENTS.md, so they behave predictably.

Evidence before claims: the model may not say work is done, fixed or passing until it has run the verification command in the same turn and read the output.

Turn a spec into a step-by-step implementation plan another engineer can follow: exact files, interfaces, test-first steps with expected output and a self-review.

Test-first development one thin slice at a time: write a failing test, make it pass, and test only at seams you agreed on.