An agent is a specialist Claude hands a whole job to, with its own tools and its own context.
1,494 agents across 238 plugins.
Use for hard problems only — cross-cutting design, gnarly multi-file debugging, race conditions and concurrency, performance investigations, database…
Use for standard implementation work — features, bug fixes, writing tests, refactors scoped to roughly 5 files or fewer, API endpoints, UI components. The…
Use proactively for mechanical, low-risk edits — renames, typo fixes, comment and docstring updates, formatting, simple config tweaks, adding log lines. Only…
Dispatch when a Go service needs a clean-architecture audit. Walks the module's import graph and flags dependency-direction violations (delivery importing…
Dispatch when a Go pull request, diff, patch, or set of staged changes needs a community-style review before merge. Walks the diff topic-by-topic using the…
Dispatch when a Go codebase has accumulated repeated patterns (idiomatic or anti-) worth capturing as a new gophers skill. Walks the module, surfaces repeated…
Ensure the Normandy Park website meets WCAG 2.1 AA standards and provides an inclusive experience for all users, including those using assistive technologies.
Implement secure authentication and user account management for the My Account portal, handling user registration, login, session management, and protected…
Specialized agent for developing, deploying, and managing Azure serverless applications including Azure Functions, Azure Static Web Apps, and Azure Table…
Adversarial reviewer. Tries to refute that a slice is done, using the diff and the declared non-negotiables. Read-only.
Implements one slice completely inside an assigned file tree. Reads real code before writing. Never publishes.
Decides open technical questions against real documentation and real APIs on the web, never from memory. Read-only.
Automated Android UI/integration testing specialist; the agent that drives a real Android app on a headless emulator under WSL/Linux and gates on what it…
Android native specialist for Kotlin, Java, Gradle/AGP, the Jetpack libraries, JNI/NDK, OpenGL ES and camera pipelines (Camera2, CameraX, MediaCodec,…
Review persona: security lens. Use for input-handling, ReDoS, prototype-pollution, supply-chain, publish-hygiene, GitHub Actions workflow exploitation, and…
Системный архитектор. Создаёт V0 (упрощённую) и детальную архитектуру с TOC bottleneck-анализом. Применяет Карпати Simplicity First (≤10 компонентов). Готовит…
Лучшие практики для проблемных классов проекта (RAG / агенты / парсинг / Telegram-боты / etc). Минимум 5 best practices из 2026 от FAANG / open-source.…
Браузерные e2e-тесты через Playwright (запуск из Bash) — основной путь. Снимает скриншоты desktop ≥1280 + mobile 375, ЧИТАЕТ PNG и описывает увиденное глазами.…
- Review diffs for correctness, consistency, and scope adherence - Verify tests exist and are meaningful for behavior changes - Flag security, performance, and…
Deep implementation work delegated by the lead — writing kit code, writing tests (five-exit-doors discipline), debugging failures, and the implementer…
Mechanical, low-judgment work delegated by the lead — boilerplate, scaffolding, multi-file sweeps and renames, doc formatting, test fixtures, applying an…
The independent review pass on a diff or PR — ONE holistic pass over the full change (code + tests + docs together), findings ranked Blocking/Important/Minor.…
Dispatches an already-prepared batch of gemini-swarm subtask prompt files to parallel agy (Gemini) sessions, waits for all of them, writes the run log, and…
Write the simplest working implementation. Use when executing code changes from a finalized spec.
Implements changes by writing code. Does not perform any testing
Validates changes by writing and executing unit test cases.
The `xhigh` rung of `cad-assumptions-analyzer` (codebase assumption gathering); `bin/route.mjs` picks it, not the user.
The `high` rung of `cad-assumptions-analyzer`; `bin/route.mjs` picks it, not the user.
The `high` rung of `cad-executor` (plan task execution); `bin/route.mjs` picks it, not the user.
Stage 4. Synthesizes the holistic verdict, score, and final report from all stage outputs via Opus reasoning.
Stage 3 leadership. Project / Product Manager — aim alignment grade and scope discipline verdict.
Stage 3 leadership. Senior Systems Architect — structural coherence verdict via ADR-style reasoning.
Use this agent when you need a design-level architecture review of a module, a proposed change, or an entire codebase. Evaluates module boundaries, coupling…
© 2026 Flowy · Free and open source
Built for Claude Code · Not affiliated with Anthropic