What is agent-isle CLAUDE.md?

Repository-specific instructions for Agent Isle, a macOS menu-bar app that displays AI coding-agent sessions in the MacBook notch and handles permission requests.

Use it when developing, building, testing, packaging, or installing hooks for Agent Isle.

It gives the agent the project's build commands, architecture, and setup needed for safe changes and end-to-end checks.

agent-isle CLAUDE.md — Instructions file by DevLab-Technologies

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/devlab-technologies/agent-isle/claude-md

Clone the repo

git clone --depth 1 https://github.com/DevLab-Technologies/agent-isle

Per session 1,582 This file is loaded in full into every session.

When invoked 1,582 The same file — it is already loaded in full.

Security scan B 1 finding. 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.

Model	Per session	Once invoked
Fable 5	$0.01582	$0.01582
Opus 5	$0.00791	$0.00791
Sonnet 5	$0.00316	$0.00316
Haiku 4.5	$0.00158	$0.00158

Measured 2d ago against content hash b2830e4129b3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

agent-isle CLAUDE.md scanned grade B with 1 finding 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

- `Scripts/install-hooks.sh` / `uninstall-hooks.sh` — shell scripts that edit `~/.claude/settings.json` (honors `CLAUDE_CONFIG_DIR`), pointing hooks at the in-repo `Scripts/agent-isle-hook.py`.

CLAUDE.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

What this is

Agent Isle is a native macOS menu-bar accessory app (Swift/SwiftUI/AppKit, no third-party dependencies) that renders a "Dynamic Island" in the MacBook notch to monitor AI coding-agent sessions, approve their permission requests, and chat with them. It runs as an LSUIElement (no dock icon) and force-terminates other running copies on launch so only one instance owns the event port.

Commands

swift build                     # debug build
swift test                      # run all tests
swift test --filter ChatHistoryTests   # run a single test target/case
bash Scripts/bundle.sh          # build "build/Agent Isle.app" (release; pass "debug" for a debug bundle)
open "build/Agent Isle.app"     # launch
bash Scripts/release.sh         # universal build + optional notarization + zip

To exercise permission approvals end-to-end you must run the app AND install the Claude Code hooks (bash Scripts/install-hooks.sh); monitoring works without hooks.

Architecture

Three cooperating layers, all under Sources/AgentIsle/:

UI (Notch/, Views/) — AppDelegate builds a fixed-size floating NotchWindow pinned over the notch (NotchGeometry detects the physical notch, else centers a pill). PassthroughView makes the window click-through everywhere except the island itself. IslandRootView toggles collapsed↔expanded.
State (Model/) — SessionStore is the single ObservableObject source of truth (sessions, filters, demo mode). Everything else feeds it or reads from it.
Ingest (Server/) — two independent input paths populate the store:
- Hook-free watching: IdeWatcher discovers Claude Code sessions by tailing ~/.claude/projects/ transcripts (TranscriptReader/TranscriptTailer extract activity + token totals); ExternalAgents adapts Grok (~/.grok/sessions) and Copilot (~/.copilot/history-session-state). This is read-only and always on.
- Event server: EventServer is an NWListener HTTP server on 127.0.0.1:4711. Any tool POSTs to /event. Read the module comment before touching it — see the parking model below.

Read the full file on GitHub · 58 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.

2d ago First seen · 58 lines · 1,582 tokens per session scan B b2830e4129b3

Subscribe to this mod's changes

agent-isle CLAUDE.md is an instructions file published in the GitHub repository DevLab-Technologies/agent-isle (43 stars, last pushed 2d ago), licensed MIT. It adds 1,582 tokens to every session, about $0.0079 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Data and AI Planning Testing ai-agents claude-code copilot developer-tools dynamic-island macos

notch-so-good CLAUDE.md

Instructions for deepshal99/notch-so-good, covering notch so good, tech stack, build, design context and users.

deepshal99/notch-so-good · 612 tokens

codex-island-app AGENTS.md

Instructions for Jarcis-cy/codex-island-app, covering repository guidelines, 沟通要求, 项目结构与模块划分, 构建、测试与开发命令 and 代码风格与命名规范.

Jarcis-cy/codex-island-app · 2,050 tokens

DevoxxGenieIDEAPlugin GEMINI.md

Instructions for devoxx/DevoxxGenieIDEAPlugin, a project described as: DevoxxGenie is an agentic plugin for IntelliJ IDEA that uses local LLM's (Ollama, LMStudio, GPT4All, Jan and Llama.cpp) and Cloud based LLMs to help review, test, explain your project code. Latest version now also supports Spec Driven Development…

devoxx/DevoxxGenieIDEAPlugin · 257 tokens

agent-isle CLAUDE.md