An AI system architect that designs software structure using Domain-Driven Design, a way to organize code around business concepts, and keeps project context in a Memory Bank. The Memory Bank is a set of files recording the project's product context, current work, and progress.
A project question-and-answer assistant that reads a folder of shared project notes called a memory bank. It reports whether that memory is active and uses the notes to answer with project context.
A structured code-auditing agent that checks a project across code quality, security, architecture, tests, documentation, and agent-instruction files. It ranks findings from Critical to Low and includes suggested fixes.
A coding agent that writes, debugs, and refactors code while following a project's recorded conventions and architecture rules. It uses a Memory Bank when one exists; a Memory Bank is a group of files describing project context and decisions.
A context-reading agent that loads project memory files, source code, and documentation, then turns them into a structured project summary. A Memory Bank is a set of files that records a project's goals, architecture, decisions, current focus, and progress.
A debugging agent that investigates software problems using the project's code, history, and decision records. It follows the same Memory Bank process as the coding agent when those project files exist.
A deep-reasoning assistant for complex engineering questions. It examines algorithms, architecture choices, technical options, and root causes in a structured way.
A coordinator for large software tasks involving several specialized assistants. It breaks work into smaller tasks, assigns each task, tracks dependencies, and combines the results.
A read-only codebase research agent that scans a project, compares related parts, and organizes its findings into a report. It responds in Traditional Chinese while keeping technical terms in English.
A code-review coordinator that sends the same review task to three different AI reviewers and combines their findings into one report. It groups shared findings, disagreements, unique findings, and likely false alarms.
A code-review assistant focused on security, type correctness, boundary cases, and error handling. It produces categorized review comments with file paths, line numbers, positive findings, and a confidence score.
Project instructions for GitHub Copilot and Claude Code in an AI-assisted development template. It defines a hierarchy of rules, shared project notes, and a layered software structure.
A workflow that updates CHANGELOG.md, the file that records what changed between software versions. It follows Keep a Changelog, a common structure for grouping additions, changes, removals, fixes, and security updates.
A code-review skill that examines code for quality, security, performance, and maintainability. It produces a report with strengths, improvement suggestions, locations, and scores.
A skill for applying domain-driven design, an approach that organizes software around business concepts, while keeping database code separate. It can create a standard structure for new features and check dependency directions.
A pre-commit documentation checker and updater for Git projects. Git is a system for tracking code changes, and a commit records a set of those changes.
A memory-checkpoint helper that saves the current project context, progress, decisions, and architecture notes in a Memory Bank folder. It is intended for long conversations where earlier details may be condensed or forgotten.