Ruitong Liu
Community@avatarganymede · Hanzhou, Zhejiang
Zhejinag University, Computer Science and Technology
Agent Skills by Ruitong Liu
Showing 46 vetted skills indexed across 2 GitHub repositories.
gui-review
Reviews Unity MVVM GUI code through parallel subagent and mechanical verification gates.
gui-prefab
Binds View.cs SerializeField references to Unity Prefab nodes in an MVVM GUI pipeline.
gui-draft
Generates MVVM code (Panel.lua and View.cs) for Unity GUI panels from a GUI_PLAN.md contract.
gui-plan
Plans Unity MVVM GUI requirements through interactive plan mode and produces an approved GUI_PLAN.md contract.
gui-learn
Extracts reusable lessons from GUI development sessions into a persistent knowledge base.
gui-config
Edits Excel config source tables and mirrors changes into *_data.lua files.
gui-improve
Iteratively fixes merged CRITICAL review findings and reruns dual-lane GUI verification gates.
playwright-cli
Automate browser interactions and web page testing with Playwright.
commit
Guide users through creating Conventional Commits with structured messages.
using-agent-skills
Routes tasks to the appropriate Kandev agent skills via mapping guides.
pr-fixup
Automates GitHub pull request fixup by polling CI checks and addressing review comments.
record
Records architectural decisions and synchronizes corresponding product specification updates during development conversations.
verify
Automate format, typecheck, test, and lint verification for a monorepo.
plan
Generates structured implementation plans and task files from feature specifications and codebase analysis.
code-review
Review code for security, architecture, logic, and test coverage compliance.
spec
Write and validate product feature specifications with testable requirements.
mobile-parity
Validate desktop and mobile UI parity with Playwright E2E tests.
context-engineering
Curate project context for coding and debugging sessions.
qa
Verify software features and hunt bugs through predefined checks.
release
Automate pcraft releases across npm, GitHub, and Docker via GitHub Actions.
pr
Automate pull request creation and management across GitHub, GitLab, and Azure Repos.
push
Automate Git commit and push operations with Conventional Commits.
spec-driven-development
Coordinate spec-driven development stages with TDD and subagent roles.
add-integration
Automate third-party integration setup with credential and authentication health checks.
Frequently Asked Questions About Ruitong Liu
FAQPage SchemaWhat tasks can I automate using AvatarGanymede's skills?▼
You can automate Git commits and pushes with Conventional Commits, pull request creation and fixup, monorepo verification (format, typecheck, lint, test), Playwright browser and E2E testing, npm/GitHub/Docker releases, bug reproduction and fixing, and kandev agent task, budget, and approval management.
Who are these skills designed for?▼
Software engineers and engineering teams practicing spec-driven development and TDD, QA engineers running Playwright E2E and UI parity checks, teams orchestrating kandev agents with budgets and escalations, and game developers building Unity MVVM GUI panels through the staged gui-plan to gui-learn pipeline.
How does the Atom Game GUI pipeline run in practice?▼
Start with gui-plan by supplying a panelId and requirement description; it produces a GUI_PLAN.md contract after plan-mode approval. The run then sequentially triggers gui-draft (MVVM code), gui-prefab (Prefab binding), gui-config (Excel tables), gui-review (dual-lane verdict), gui-improve (max 2 fix rounds), and gui-learn (knowledge writeback).
What prerequisites and dependencies do these skills require?▼
Playwright skills require a Playwright environment; verification skills expect a monorepo with format, typecheck, test, and lint commands; PR skills target GitHub, GitLab, or Azure Repos; kandev skills need the kandev orchestrator and agentctl; GUI pipeline skills require the Atom Game project with Claude Code plan mode.
How do the kandev skills manage agents and costs?▼
Kandev skills list workspace agents with roles and budgets, hire agents with approval policy enforcement, manage task lifecycles via agentctl, schedule recurring tasks through cron and webhook triggers, escalate blocked decisions to human review, monitor token spend per agent, and export or import .kandev/ workspace configurations.