claude-plugins

Runs an adaptive, project-aware code-review workflow (profile the project, detect its lint tools + conventions, run mechanical + semantic review, consolidate into an LGTM verdict) and carries the Google eng-practices playbook it applies — reviewer (what to look for, when to approve, severity-labeled comments, handling pushback) and PR/CL author (CL descriptions, splitting changelists, responding to comments). Use when reviewing or authoring a PR/MR/CL/diff, running a code review, deciding LGTM or LGTM threshold, authoring a merge request, assessing code health, handling a hotfix review, or replying to review feedback.

Push books/docs and read notes from an Onyx BOOX e-reader account (push.boox.com / send2boox.com) WITHOUT the web UI, via the self-contained `boox.mjs` CLI. Uploads a file to Aliyun OSS, writes a digital_content doc into the device's MESSAGE sync channel, and calls saveAndPush to notify the device — the exact 3-step flow the web app performs. Also lists the push list, lists synced notes, removes pushed items, and shows device online status. Auth = long-lived JWT via BOOX_TOKEN env or a token file.

Capture this session's repeatable process into a reusable SKILL.md file via guided interview.

Turn Claude into a wise, effective teacher whose only goal is to make sure YOU deeply understand the work an AI agent just did — the problem it solved, why that problem existed, the solution and its design decisions, the edge cases it handled, and the broader impact. Incremental and mastery-gated: it has you restate your understanding first, fills the gaps at the depth you ask for (eli5 / eli14 / like-an-intern), shows you the real code, and quizzes you with AskUserQuestion — and it does not conclude until you have demonstrably understood. Use when the AI moved faster than you could follow and you want to catch up before merging. Trigger on "teach me", "help me understand what you did", "make sure I understand this change", "walk me through this", "explain the session", "I don't get it", "why did you do it this way", "quiz me", "eli5 / eli14", "don't let me fall behind", "I want to actually understand this, not just merge it", "catch me up", "bring me up to speed", "I fell behind", "the AI moved too fast".

Architecture decision advisor — guides pattern selection, application, synthesis, and stress-testing for system design, scaling, resilience, and migration decisions.

Look up architecture patterns and concepts via semantic search — read the full article on demand, compare patterns side-by-side, or browse by volume.

Decision and reasoning advisor — picks 2-3 relevant mental models, walks through applying each, and synthesizes a recommendation. Use proactively when facing a hard trade-off, an ambiguous decision, or any "should we do X or Y" question where structured reasoning beats gut feeling.

Threat modeling framework advisor — selects the right framework(s) for the system (security, privacy, AI/ML, cloud, K8s), walks through application, and composes findings across frameworks. Use proactively when reviewing architecture for risk, conducting a security architecture review, assessing a system's attack surface, enumerating threats, analyzing security posture, modeling vulnerabilities, or performing a structured risk assessment.

Extract cookies from local browsers and export them for use in authenticated requests. Use when extracting Chrome cookies, Firefox cookies, or Safari cookies from the local browser store; reusing a browser session to replay an authenticated request; exporting a Netscape cookie jar for curl, wget, or yt-dlp cookies; making an authenticated curl request with your live browser session; or using rookiepy to read browser cookies programmatically. Also covers session replay — sending HTTP requests with the browser's own cookies injected automatically.

Bootstrap or scaffold a Bun fullstack project in one shot — single-port server, Vite dev proxy, monorepo layout, and Docker config. Use when starting a new fullstack project with Bun, setting up a bun server with vite proxy, creating a monorepo bun layout, configuring docker bun deployment, or bootstrapping a bun project from scratch. Skips files that already exist so it never overwrites your work.

Submit bug reports, feature requests, and plugin requests to aiocean/claude-plugins via GitHub Issues.

Generate comprehensive architecture documentation — the Codebase Oracle that writes docs, module docs, and dependency maps — powered by the GitNexus knowledge graph and LSP analysis.

Install, configure, and manage the GitNexus code intelligence engine — index codebase, setup MCP, check status, troubleshoot, and document local git-hook auto-refresh for master-only workflows.

Synthesize discovery and dependency mapping into a step-by-step implementation plan via planner agent. Step 3 of the codeflow discover → map → plan trio (after aio-discover and aio-map).

Quick pre-commit sanity check — validate post-implementation changes against the aio-snapshot baseline via change detection, duplication detection, and type checking. Light pass; for deep multi-agent review use aio-review-deep.

Web UI design and review advisor — applies UX psychology, interaction patterns, WCAG accessibility, and modern CSS to design new interfaces or critique existing ones. Use proactively when reviewing any frontend UI, scaffolding new components, or deciding layout/color/typography trade-offs.

Read and manage GitLab merge requests, CI/CD pipelines, branches, and jobs via gitlab-mcp (auto-installs if missing). Use when the user mentions a GitLab URL, asks to open or review an MR, debug a failing pipeline, retry a CI job, protect a branch, trigger a pipeline run, or run any GitLab operation programmatically via glab or the MCP server. Supports gitflow commands (create feature, finish release, hotfix) and self-hosted GitLab instances.

Manage git worktrees for parallel development workflows — create, sync, and spotlight preview.

Create Grafana diagrams for live system visualization — analyzes codebase to auto-generate Mermaid diagrams with Prometheus metric binding. Each node is colored by health thresholds, turning a static architecture diagram into a live system diagram or service health monitor. Use when you need a dynamic Grafana dashboard with node coloring driven by real metrics. For standalone shareable diagrams use aio-mermaid instead.

Generate shareable Mermaid diagram URLs (LZ-String compressed). Use when you want to draw a diagram, create an architecture diagram, generate a flowchart, make a sequence diagram URL, build a mindmap, timeline, or gitgraph — and share it as a link. Supports all Mermaid v11 types: flowchart, sequenceDiagram, classDiagram, stateDiagram-v2, erDiagram, gantt, pie, mindmap, timeline, gitgraph.

Pre-translation analysis for EPUB books — analyze writing style, characters, tone, and build glossary before translating.

Pack a translated EPUB into a downloadable file — bilingual (original + translation side-by-side) or translation-only output. Use when the user wants to export/xuất sách, đóng gói sách, generate an EPUB file, download a finished book, or send a translation to Kindle.

Browse books, check translation progress, manage guidelines, view TOC, fork books, publish to community, reset chapters, and view usage stats.

Generate a translation quality report and surface chapters that need re-translation. Use when the user wants to check quality, kiểm tra chất lượng bản dịch, run QA, tìm bản dịch kém, find bad translations, or decide which chapters to retranslate before exporting.

Deep literary review of translated EPUB chapters — evaluate translation quality, fix consistency issues, and check cross-chapter coherence.

Translate EPUB book chapters into literary Vietnamese with cross-chapter consistency (guideline + glossary + previous-chapter summary). Use when the user says dịch sách, dịch chương, dịch tiếp, continue translating, translate epub, or sends a book/chapter URL to translate.

Upload an EPUB file and prepare it for chapter translation — extracts TOC, counts words, queues for analysis. Use when the user wants to add a new book, start translating an EPUB, tải sách lên, import a `.epub` file, or onboard a new title into the translation workspace.

Vietnamese literary translation style guide — word choice, sentence rhythm, genre conventions, and common error patterns for natural Anh-Viet translation.

Write, review, and lint Go code. Lint mode runs go build, go vet, golangci-lint, govulncheck, nilaway, deadcode, and race detection (race detector), then applies idiomatic fixes. Reference mode covers concurrency, error handling, generics, testing, gRPC, and production hardening. Use when asked to lint golang, run a go lint pipeline, review go code quality, do go static analysis, write idiomatic go, or apply go best practices.

Remove image backgrounds via RMBG-2.0 alpha matting with despill + 1px erode — produces clean PNG RGBA cutout that handles hair, smoke, glow, soft edges. Use when the user wants to remove background, xoá nền, make transparent, cutout, chroma key, or post-process a text-to-image render (gpt-image, Imagen, FLUX, SDXL, Grok, Midjourney) — especially with flat magenta (#FF00FF) key.

Debug iOS apps on physical devices from the terminal — deploy to iPhone or iPad with xcodebuild, capture syslog via idevicesyslog, pull crash reports from the device, analyze .ips crash files, and take device screenshots. Covers the full loop: build, install, launch, capture syslog, pull crash report, analyze .ips crash file, and grab a device screenshot. Uses libimobiledevice (idevicesyslog, idevicecrashreport) and pymobiledevice3 for iOS 17+ compatibility. Use when debugging an iOS app on a physical device, deploying to iPhone without the Xcode GUI, troubleshooting a crash from a .ips file, or capturing a device screenshot over USB.

Watch the kanban board (`.kanban/board.md`) and auto-dispatch a sub-agent for each task in a chosen column whose content matches a free-text goal. Use when you want one Claude session to run "headless executor" while another session triages and writes new tasks into the board. Trigger phrases: "monitor kanban", "auto run tasks", "watch board", "auto dispatch tasks", "kanban worker", "kanban auto", "/aio-kanban-monitor".

Markdown-based kanban board for AI agent task management. Board is a pure index (`.kanban/board.md`), every task lives in its own file (`.kanban/tasks/T-NNN-slug.md`). Use to init the board, add tasks, move them across Backlog/Todo/Doing/Done/Blocked, or archive completed work.

Bridge Claude to a browser UI in real time via a frozen Bun + Vue3 + Tailwind scaffold — browser events become Monitor-tool notifications, AI pushes become WebSocket broadcasts. Use when Claude needs to drive an interactive UI mid-task: form capture, multi-step decision flow, live preview, approval queue, or side-by-side review. See body for the "why" (Claude has no event loop) and architecture details.

Give Claude Code an event loop. A local HTTP + WebSocket relay plus the Monitor tool let anything OUTSIDE Claude — a CLI in another terminal, a mobile app, a webhook, another running program, a web page — push events to Claude mid-task and receive pushes back, in real time. Inbound: the external client POSTs an event, the relay prints a MSG:: line to stdout, the Monitor tool surfaces it as a notification. Outbound: Claude POSTs a push, the relay broadcasts it over WebSocket to every connected client. The protocol is generic and language-agnostic; a ready-to-run reference relay is included. Use when an external process or device needs a live two-way channel to Claude, when you need a webhook or event source to wake Claude, or when you want to "give Claude an event loop". "external to Claude bridge", "Monitor + WebSocket relay", "wake Claude from outside", "kết nối bên ngoài với Claude", "cầu nối tương tác với Claude".

Design metrics, alerts, dashboards, and SLOs for any service using monitoring and observability best practices. Use when setting up observability from scratch, designing alert rules, building Grafana dashboards, integrating OpenTelemetry instrumentation, configuring Prometheus, implementing SLI/SLO tracking and error budgets, aggregating logs with ELK or Loki, debugging with distributed tracing, applying the RED method or USE method or Four Golden Signals, migrating from Datadog to an open-source stack, or comparing monitoring tools like Datadog, Grafana Cloud, Jaeger, and Tempo.

Set up vector database collections, index content, and run semantic search via RAG pipelines using Qdrant and rag-kit (auto-installs if missing). Use for retrieval-augmented generation, similarity search over a knowledge base, or embedding content for meaning-based retrieval.

Run a structured 10-phase research workflow to investigate a topic, conduct a literature review, or deep-research a technical or market question via research-cli (auto-installs if missing). Use for any investigation that needs phase-gated discipline: question formulation → data collection → analysis → documentation.

Read, create, and update Jira issues, sprints, and workflows plus Confluence pages via the Atlassian CLI. Use when the user mentions a Jira ticket key (e.g. PROJ-123), asks for a JQL/CQL search, wants to transition an issue, manage a sprint/backlog, or read/write a Confluence page.

Interact with Google Workspace services via the gws CLI. Use when the user wants to send email, read Gmail, upload to Drive, create a Google Calendar event, read or write Google Sheets, generate a standup report, manage Google Tasks, or work with Google Docs, Slides, Chat, or Contacts.