Andrew Hundt ahundt

Claude Code + Serena MCP Auto-Wrapper

Zero-configuration automatic Serena MCP server management for Claude Code

Transparently starts exactly one Serena instance per project with unique ports. No per-project setup required!

Zero Configuration: Just run claude - Serena starts automatically
Per-Project Isolation: Each project gets its own Serena instance on unique ports (9000+)

Open Automator
Create a new document
Select Quick Action
Set “Service receives selected” to files or folders in any application
Add a Run Shell Script action
- your default shell should already be selected, otherwise use /bin/zsh for macOS 10.15 (”Catalina”) or later
- older versions of macOS use /bin/bash
if you're using something else, you probably know what to do 😉

First download the new old icon: https://cl.ly/mzTc (based on this)

You can also use the icon you want, but you need to convert it to .icns. You can use this service to convert PNG to ICNS.

Go to Applications and find VSCode, right click there and choose Get Info. Drag 'n drop the new icon.

	; このファイルを `~/.claude/claude-sandbox.sb` に設定したあと、
	; alias c="sandbox-exec -f ~/.claude/claude-sandbox.sb -D PWD="`pwd`" -D HOME="$HOME" claude "
	; このaliasを~/.zshrcに書き込んで、`. ~/.zshrc`で読み直します。
	; dangerousな人は下記のように。
	; alias c="sandbox-exec -f ~/.claude/claude-sandbox.sb -D PWD="`pwd`" -D HOME="$HOME" claude --dangerously-skip-permissions "
	; これで `c`だけで安全にClaude Code CLIが実行できます。

	(version 1)

	; based

	# Project Policy

	This policy provides a single, authoritative, and machine-readable source of truth for AI coding agents and humans, ensuring that all work is governed by clear, unambiguous rules and workflows. It aims to eliminate ambiguity, reduce supervision needs, and facilitate automation while maintaining accountability and compliance with best practices.

	# 1. Introduction

	> Rationale: Sets the context, actors, and compliance requirements for the policy, ensuring all participants understand their roles and responsibilities.

	## 1.1 Actors

	class weighted_categorical_crossentropy:
	def __init__(self, weights):
	self.weights = weights
	self.__name__ = 'wcentroid_loss'

	def __call__(self, y_true, y_pred):
	class0 = K.ones_like(y_pred)[:, :, :, 0] * self.weights[0]
	class1 = K.ones_like(y_pred)[:, :, :, 0] * self.weights[1]
	x = K.tf.where(y_true[:, :, :, 0] > 0, class0, class1)
	result = x * K.categorical_crossentropy(y_pred, y_true)

	import datetime as dt

	import tensorflow as tf
	import tensorflow.contrib.slim as slim
	from tensorflow.contrib.slim.nets import resnet_v1
	import threading

	from PoseDataset import PoseDataset
	from TrainParams import TrainParams

	import tensorflow as tf
	from tensorflow.python.framework import ops
	import numpy as np

	# Define custom py_func which takes also a grad op as argument:
	def py_func(func, inp, Tout, stateful=True, name=None, grad=None):

	# Need to generate a unique name to avoid duplicates:
	rnd_name = 'PyFuncGrad' + str(np.random.randint(0, 1E+8))