SubFrame CLAUDE.md

Project instructions for SubFrame, an Electron desktop app that gives Claude Code a terminal-based interface. They describe its architecture, technology choices, and rule that agents should make only requested changes.

In plain words
What is it for?
They guide work on the React interface, Node.js process, typed communication between them, file handling, terminal sessions, state management, and build-related modules.
Why use it?
They help coding assistants understand where the app's user interface, background process, shared communication code, and key modules belong.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/codename-11/subframe/claude-md
Clone the repo
git clone --depth 1 https://github.com/Codename-11/SubFrame
Per session 5,031 This file is loaded in full into every session.
When invoked 5,031 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.05031 $0.05031
Opus 5 $0.02516 $0.02516
Sonnet 5 $0.01006 $0.01006
Haiku 4.5 $0.00503 $0.00503

Measured 2d ago against content hash e940e7ee0a54, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

SubFrame CLAUDE.md 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 2d 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.

CLAUDE.md · 244 lines

How it starts

The opening of the file, as written. The whole thing — 244 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SubFrame

Terminal-centric IDE for Claude Code. Enhances native AI tools — does not replace them.

Core Rule

Only do what the user asks. Do not refactor, restructure, or add features beyond the request. Present suggestions separately — never implement without approval.

Architecture

Electron app with two processes communicating over typed IPC (src/shared/ipcChannels.ts). The renderer uses a pluggable Transport interface (src/shared/transport.ts) — currently ElectronTransport, with WebSocketTransport planned for browser/mobile access (ADR-006).

Layer Entry Point Purpose
Main (Node.js/TS) src/main/index.ts PTY management, file I/O, IPC handlers, manager modules
Renderer (React/TS) src/renderer/index.tsx React 19 + Zustand + TanStack Query, shadcn/ui, Tailwind CSS v4
Shared src/shared/*.ts Typed IPC channels, constants, templates, backlink utilities

Tech Stack

React 19, TypeScript (strict), Zustand (state), TanStack Query (IPC caching), TanStack Table (sorting/filtering), shadcn/ui (components), Tailwind CSS v4, Framer Motion (animations), CodeMirror 6 (editor), esbuild (bundler).

Key Modules

Main process — each manager has init() + setupIPC(): ptyManager tasksManager pluginsManager claudeSessionsManager aiToolManager aiFilesManager settingsManager gitBranchesManager overviewManager agentStateManager skillsManager promptsManager updaterManager sessionSnapshotManager pipelineManager activityManager outputChannelManager onboardingManager popoutManager claudeUsageManager githubManager apiServerManager webServerManager workspace frameProject Main utilities (no manager pattern): fileEditor (IPC file read/write) dialogs (native OS dialogs) menu (Electron application menu) promptLogger (prompt history logging) pty (PTY spawn helpers) ipcRouter (dual-registration IPC for Electron + WebSocket) eventBridge (broadcast fan-out to Electron window + WS clients) aiSessionManager (background AI session lifecycle — PTY spawn, prompt injection, result extraction) aiExecutionManager (per-tool CLI flag construction, invocation modes, interactive tool configs) Shared utilities: taskMarkdownParser (parse/serialize task .md files with YAML frontmatter) pipelineWorkflowParser (YAML workflow parsing) pipelineStages (built-in stage handlers) Shared types: themeTypes (theme tokens, presets, CSS mapping) activityTypes (activity stream types, IPC events) agentStateTypes (agent session/step types) subframeHealth (health check types) transport (Transport + TransportPlatform interfaces — pluggable IPC abstraction) Shared helpers: backlinkUtils (CLAUDE.md/GEMINI.md backlink injection) claudeSettingsUtils (settings.json merge) projectInit (workspace init logic) logoSVG (inline SVG logo) frameConstants (version, paths) frameTemplates (file templates for init)

Read the full file on GitHub · 244 lines

Changes

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.

  1. 2d ago First seen · 244 lines · 5,031 tokens per session scan A e940e7ee0a54

Subscribe to this mod's changes

SubFrame CLAUDE.md is an instructions file published in the GitHub repository Codename-11/SubFrame (10 stars, last pushed 8d ago), licensed MIT. It adds 5,031 tokens to every session, about $0.0252 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.