A ready-to-use, project-customizable AI Coding Agent Skills framework providing general Workflows, extensible project-specific knowledge, a self-learning loop, and troubleshooting capabilities, compatible with any tech stack and engineering scenario. Compatible with mainstream Coding Agents.
A read-only investigation of a software codebase, including its modules, dependencies, call paths, configuration, and tests. The report distinguishes direct evidence from conclusions drawn from it.
A systems-code review agent that tracks every changed decision branch from its contract through the evidence needed to trust it. A branch is a path such as an early return, fallback, allow-list check, or null check.
An adversarial review agent that tries to disprove specific software-review findings submitted by another agent. It returns whether each finding was rejected, not proven, or could not be rebutted.
A code-review agent focused only on reliability in unusual situations, including boundary cases, error handling, resource cleanup, concurrency, and exception paths. Its instructions are written in Chinese.
A software-review agent that checks whether code matches the requested features, design documents, tasks, and interface agreements. It focuses on expected behavior and missing requirements.
A project-rules reviewer that checks code against explicitly stated engineering standards, such as naming, file structure, dependencies, logging, and synchronization rules.
★not rated 159 7mo agoA46 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: