max-patch-agent

A specialist assistant for generating MAX patches, which are visual programs built by connecting objects. It focuses on control and data flow, message routing, subpatchers, MIDI, and user-interface objects.

In plain words
What is it for?
Use it to construct MAX control logic, route messages, organize subpatchers, handle MIDI, and generate validated .maxpat files.
Why use it?
It creates patches using the project's object database and checks package restrictions and supported objects while building them.

Skill for Claude CodeCodex

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 skills/taylorbrook/max-msp_cc_framework/max-patch-agent
Any agent
npx skills add taylorbrook/MAX-MSP_CC_Framework --skill max-patch-agent
Clone the repo
git clone --depth 1 https://github.com/taylorbrook/MAX-MSP_CC_Framework

Made for: Claude Code, Codex.

Per session 23 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,153 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00023 $0.05153
Opus 5 $0.00012 $0.02576
Sonnet 5 $0.00005 $0.01031
Haiku 4.5 $0.00002 $0.00515

Measured 2d ago against content hash 8be8a75c2f0c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

max-patch-agent 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/skills/max-patch-agent/SKILL.md · 348 lines

How it starts

The opening of the file, as written. The whole thing — 348 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Patch Generation Specialist

The Patch agent generates MAX/MSP .maxpat files focused on control-rate operations: object routing, message passing, subpatcher organization, MIDI handling, and data management.

Domain Context Loading

Before any generation:

  1. Read CLAUDE.md at project root -- follow all 5 rules and patch style guidelines
  2. Use ObjectDatabase from src.maxpat.db_lookup for all object lookups -- it loads all domains, resolves aliases, checks PD blocklist, and provides relationship data automatically
  3. Read .claude/max-objects/relationships.json for common object pairings (if needed for design decisions)
  4. Read project config.json via load_project_config() from src.maxpat.project for allowed packages. Pass allowed_packages to Patcher(allowed_packages=allowed) so package objects outside the project's selection are blocked at creation time.

Domain focus: Max control/data/UI objects. Signal processing and RNBO are handled by their respective agents.

Capabilities

Patch Construction

  • Create Patcher instances with boxes and connections via src.maxpat.patcher
  • Use Box constructor for all standard objects (validates against ObjectDatabase)
  • Use Box.__new__() bypass for structural objects: subpatchers, bpatcher
  • Connect boxes with Patcher.add_connection(src_box, src_outlet, dst_box, dst_inlet)

Key Functions

  • Patcher() -- create a new patch
  • Box(name, args, db) -- create a validated box
  • Patcher.add_box(box) -- add box to patch
  • Patcher.add_connection(src_box, src_outlet, dst_box, dst_inlet) -- connect boxes
  • Patcher.add_subpatcher(name, inlets, outlets, inlet_comments, outlet_comments) -- add a subpatcher with labeled I/O
  • Patcher.populate_assistance_comments() -- auto-fill empty inlet/outlet comments from connection context
  • finalize_patch(patcher, is_new=True) -- single-call layout cleanup: styling, layout, comments, midpoints (new); midpoints + comments (edit)
  • apply_layout(patcher, layout_options=None) -- row-based topological layout positioning (accepts LayoutOptions)
  • validate_patch(patcher.to_dict(), db=patcher.db) -- run four-layer validation pipeline
  • save_patch_roundtrip(patcher.to_dict(), path) -- write .maxpat to disk
  • Patcher.add_comment(text, x, y) -- add a comment box (for inline annotations, critic notes)
  • Patcher.add_message(text, x, y) -- add a message box (for triggering messages, storing values)
  • Patcher.add_node_script(filename, code=None, num_outlets=2, x, y) -- add a node.script box for Node for Max (returns tuple of Box and code string)
  • Patcher.add_js(filename, code=None, num_inlets=1, num_outlets=1, x, y) -- add a js object box for V8 JavaScript (returns tuple of Box and code string)

Read the full file on GitHub · 348 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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 · 348 lines · 23 tokens per session scan A 8be8a75c2f0c

Subscribe to this mod's changes

max-patch-agent is a skill published in the GitHub repository taylorbrook/MAX-MSP_CC_Framework (41 stars, last pushed 6d ago), licensed MIT. It adds 23 tokens to every session and 5,153 once invoked, about $0.0001 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 2026-08-30.

Related

Other skills, from other repositories

html-ppt-zhangzara-8-bit-orbit

A gamer's journey building a retro-arcade collection — the obsession, the hunt, and what the machines came to mean. Built as a decision-grade story deck for friends, hobby community.

nexu-io/open-design · 53 tokens

game-art

Game art principles. Visual style selection, asset pipeline, animation workflow.

vudovn/ag-kit · 17 tokens

asset-pipeline

Generate game-ready 3D assets for World of ClaudeCraft with the AI asset pipeline (Tripo API + optional gpt-image-2). Use when asked to create or generate game assets, a new weapon model, a prop, a creature or mob model, a player-class skin, or any 3D model or texture for the game. Drives…

levy-street/world-of-claudecraft · 0 tokens

hatch-pet-v1

Create, repair, validate, preview, and package legacy Codex v1 animated pet spritesheets from character art, screenshots, generated images, or visual references. Use when a project requires the 8x9, 1536x1872 pet contract with transparent unused cells, row-by-row animation prompts, QA contact sheets, preview videos…

legeling/awesome-codex-pet · 107 tokens

art-bible

Guided, section-by-section Art Bible authoring. Creates the visual identity specification that gates all asset production. Run after /brainstorm is approved and before /map-systems or any GDD authoring begins.

Donchitos/Claude-Code-Game-Studios · 47 tokens

muapi-ai-fight-scene

Generate a high-cut-density action / fight scene by first composing a 16-cell storyboard image, then driving Seedance 2.0 image-to-video off that storyboard. Stacks GPT-Image-2 (character sheet + storyboard), Nano-Banana-2 (environment concept), and Seedance 2.0 i2v.

SamurAIGPT/Generative-Media-Skills · 75 tokens