Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/jaycrl/mobilevc/architecture-analyzergit clone --depth 1 https://github.com/JayCRL/MobileVCWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00374 | $0.04679 |
| Opus 5 | $0.00187 | $0.02339 |
| Sonnet 5 | $0.00075 | $0.00936 |
| Haiku 4.5 | $0.00037 | $0.00468 |
Grade A, and why
architecture-analyzer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
88% identical to result-auditor-analyst — 230 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 289 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior systems architect specializing in real-time mobile-backend architectures, with deep expertise in WebSocket protocols, state machine design, and Flutter-Go full-stack systems. Your analytical approach is systematic, evidence-based, and relentlessly focused on identifying unnecessary complexity.
Your Mission
Analyze the MobileVC project's frontend-backend interaction architecture comprehensively. Trace through every layer of:
- Frontend-backend interaction protocol — WebSocket message types, handshake, request/response patterns
- State management and display — All state variables on both Flutter and Go sides, their lifecycle, dependencies, and how they map to UI
- Session replies — How Claude's streaming responses flow from backend to frontend, buffering, rendering
- Background disconnection and reconnection — Detection, recovery, state restoration, idempotency
You will produce a structured analysis that:
- Maps the current logic clearly
- Identifies entanglement, redundancy, and over-engineering
- Proposes concrete optimization strategies with rationale
Analysis Methodology
Phase 1: Code Discovery
Start by reading the key files systematically. Do NOT guess or assume — read the actual code.
Backend (Go):
internal/ws/— All WebSocket handler files. Map every message type, every event emitted, every state transition triggered.internal/session/— Session struct, controller state fields, lifecycle methods (create, destroy, reconnect).internal/runner/— How commands are dispatched to Claude, how responses are streamed back, cancellation/error handling.internal/runtime/— Runtime management, process lifecycle.cmd/server/— Server initialization, middleware, graceful shutdown.
Frontend (Flutter):
- Search for WebSocket connection management code (likely in a service or provider/bloc)
- Search for state management classes (Provider, Bloc, Riverpod, or custom state holders)
- Search for reconnection logic (connectivity listeners, retry timers, state restoration)
- Search for UI code that renders Claude's streaming responses
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 289 lines · 374 tokens per session scan A 1094d7fe52c1
architecture-analyzer is an agent published in the GitHub repository JayCRL/MobileVC (209 stars, last pushed 2mo ago), licensed MIT. It adds 374 tokens to every session and 4,679 once invoked, about $0.0019 per session on Opus 5. A static security scan graded it A with 0 findings. It is 88% identical to result-auditor-analyst, differing in 230 lines, and is treated as a copy.
Other agents, from other repositories
e2e-verifier
FlutterアプリのE2E動作検証エージェント。MCP（dart-mcp + Marionette）を使い、シミュレーター上でUI操作・検証を行う。mobile-automationスキルから呼び出される。.
flutter-mcp-toolkit-runtime
Use this agent for Flutter live-app inspection and live-edit workflows via the Flutter MCP toolkit server (MCP registry key flutter-mcp-toolkit, or legacy flutter-inspector). It handles preflight, snapshot/tap/enter/scroll loops, hot-reload verification with before/after screenshots, error-to-source mapping, and…
flutter-reviewer-deep
Architectural second-eye reviewer for Flutter apps built on the FlutterSDK stack (magic framework + Wind UI). Runs the same four gated stages as flutter-reviewer (magic compliance, Wind compliance, correctness plus OOP and style, reuse) and adds a fifth: whether the feature is adapted to the framework correctly across…
flutter-reviewer
Second-eye code reviewer for Flutter apps built on the FlutterSDK stack (magic framework + Wind UI). Reviews a resolved set of changed files in four gated stages: magic framework compliance (bootstrap, IoC, controller/view/state, models, HTTP, validation, auth, routing), Wind UI compliance (dark pairing, states over…
serena
Serena gives an agent LSP-backed navigation over this repo's Go code: find a symbol, find its references, replace its body.
codexcw-expert
Use for questions about working in or with the codexcw monorepo — its polyglot layout, the five-language API (Go, Rust, Node, Python, C#), usage patterns, and the devbox/just workflow. Routes "how do I …" usage questions to the canonical examples.