Plugin Claude Code
Plugin marketplace listing 1 plugin: claude-harness.
Plugin Claude Code
Plugin marketplace listing 1 plugin: claude-harness.
Plugin Claude Code
A general-purpose harness for an AI-driven software development team.
Agent
A skeptical verification agent for claims about dependencies shared across code repositories. It independently retrieves the dependency's source code and tries to disprove the stated assumptions and evidence.
Agent
A review agent that questions decisions to dismiss comments on a pull request. A pull request is a proposed code change submitted for review.
Agent
An agent that reviews source code and implementation changes. Its description is in Japanese and indicates that it is assigned for code, implementation, and pull-request reviews.
Agent
An agent that scans a project directory for technical debt, meaning code or design choices that may make future changes harder or riskier.
Agent
An agent that checks whether reported technical debt is genuinely a maintenance problem. It reads the affected files and questions findings made by another scanner.
Agent
A judging agent that scores three proposed ways to split an implementation task, combines their best parts, and produces one final plan.
Agent
A review agent that checks an entire code change for one specified class of defects. It works independently from the implementation and only finds and reports problems.
Agent
A review agent that checks whether an implementation follows the important design decisions recorded in its feature requirements.
Agent
A design-review agent that examines software architecture, dependencies, and other design decisions.
Agent
An agent that checks whether the implementation matches the project’s requirements and design documents. It compares the written plans with the code that was built.
Agent
An agent for implementing end-to-end tests, which test a complete user flow through an application. Its description is in Japanese and specifies automatic delegation for requests to write or add such tests.
Agent
A review agent for explanations of completed end-to-end tests. End-to-end tests check a complete user flow through an application.
Agent
An agent that checks whether end-to-end tests can detect a real defect by adding one intentional bug to an important user flow. End-to-end tests exercise a complete flow through the application.
Agent
An implementation agent for adding features, changing existing behaviour, and fixing bugs, based on requests written in Japanese.
Agent
A code-review checker that tests whether reported problems are real by comparing the review comment with the changed code and the surrounding code.
Agent
An agent that predicts file conflicts and dependencies between multiple issues being implemented at the same time. An issue is a tracked piece of development work, such as a bug or feature request.
Agent
An agent that sorts one pull-request review comment into an appropriate response or follow-up action.
Agent
An agent that critiques feature specification documents for testable acceptance criteria, internal consistency, and usefulness for later implementation.
Agent
An agent for mechanically correcting blockers identified in a feature specification document.
Agent
A review agent that reads test code and identifies the behavior each test actually checks, separating public behavior from internal implementation details.
Agent
An agent that turns a parent requirements ticket into a proposed set of implementation tasks.
Agent
A worker agent that implements one issue from start to finish inside a worktree, which is a separate working copy of a code repository.