What is sanzaru CLAUDE.md?

A repository guide for Claude Code, an AI coding assistant, explaining how to develop and run a server that connects to AI services for creating videos and images. It documents setup commands, architecture, security rules, and testing practices.

Use it when installing dependencies, running the server in local or web-connected modes, changing its video or image workflows, or running checks.

It gives the coding assistant the project context and working rules needed to make safer, more consistent changes.

How TJC-LP/sanzaru configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to TJC-LP/sanzaru. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/TJC-LP/sanzaru/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/TJC-LP/sanzaru

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/tjc-lp/sanzaru/claude-md/github.svg)](https://agentmods.dev/instructions/tjc-lp/sanzaru/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/tjc-lp/sanzaru/claude-md"><img src="https://agentmods.dev/badge/instructions/tjc-lp/sanzaru/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/tjc-lp/sanzaru/claude-md"><img src="https://agentmods.dev/badge/instructions/tjc-lp/sanzaru/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 19,838 This file is loaded in full into every session.

When invoked 19,838 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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 5.1	$0.19838	$0.19838
Opus 5.5	$0.07935	$0.07935
Sonnet 5.5	$0.03968	$0.03968
Haiku 4.5	$0.01984	$0.01984

Measured 1mo ago against content hash 594150b83ad5, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

sanzaru 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 1mo 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 · 1,265 lines

How it starts

The opening of the file, as written. The whole thing — 1,265 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.

Project Overview

A stateless MCP server (mcp.server.mcpserver.MCPServer, mcp SDK 2.x, protocol 2026-07-28) wrapping Higgsfield's video API (Seedance 2.5, Kling 3.0) and OpenAI's image, speech and transcription APIs. OpenAI removed the Sora Videos API on 2026-09-24; 0.13.0 moved video generation to Higgsfield. Supports both stdio (for MCP clients) and HTTP streaming (for web clients) transports. Exposes MCP tools for async video/image generation with polling-based workflows.

Key Architecture Principles:

Stateless: No database, no in-memory job tracking. Job state lives with the provider (Higgsfield for video, OpenAI for images).
Async job pattern: Create → wait_for → Download for both videos and images (the wait runs server-side)
Security sandbox: Reference images restricted to configured media paths with path traversal protection
Type-safe: Extensive use of TypedDict and Literal types from OpenAI SDK
Dual transport: stdio (default) for Claude Desktop, HTTP for web clients and remote access

Development Commands

# Install dependencies (CI uses --all-extras; the dev group already includes the mcp extra)
uv sync

# Run the MCP server (stdio mode - default)
uv run sanzaru
uv run sanzaru serve            # explicit alias

# Run the MCP server (HTTP mode - stateless). Loopback runs open; any other bind
# refuses to start without a bearer token (or the proxy escape hatch — see Transport Modes)
uv run sanzaru --transport http
uv run sanzaru --transport http --port 3000
SANZARU_HTTP_TOKEN="$(openssl rand -hex 32)" uv run sanzaru --transport http --host 0.0.0.0 --port 8080

# Agent CLI (same tools as shell commands — see docs/cli.md)
uv run sanzaru capabilities
uv run sanzaru video create "a cat stretches" --duration 4 --resolution 480p --max-cost 1 -o ./out/cat.mp4
uv run sanzaru image generate "an icon" --quality high -o ./art/icon.png
uv run sanzaru podcast generate @script.json --verify -o ep.mp3   # check the audio says it
uv run sanzaru podcast rundown "why TTS drops sentence tails" --acts 3 -m 6 -o rundown.json
uv run sanzaru podcast simulate @rundown.json --dry-run          # plan + cost, spends nothing
uv run sanzaru podcast simulate @rundown.json --max-cost 2.00 -o ep.mp3

# Lint and format code
ruff check .
ruff format .

# Test the server locally (requires Claude Code or MCP client)
claude  # in this directory with .mcp.json configured

Read the full file on GitHub · 1,265 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.

1mo ago Changed · +8 lines · +223 tokens per session 594150b83ad5
1mo ago Changed · +54 lines · +1,650 tokens per session 5cefd09d553b
1mo ago Changed · +111 lines · +1,925 tokens per session bc0885c23e3f
1mo ago Changed · +28 lines · +403 tokens per session bfbb048ce811
1mo ago Changed · +10 lines · +202 tokens per session 54881032f3a6
1mo ago Changed · +255 lines · +5,528 tokens per session e662183be2b5
2mo ago First seen · 799 lines · 9,907 tokens per session scan A 1a846ca1ef4b

Subscribe to this mod's changes

sanzaru CLAUDE.md is an instructions file published in the GitHub repository TJC-LP/sanzaru (6 stars, last pushed yesterday), licensed MIT. It adds 19,838 tokens to every session, about $0.0794 per session on Opus 5.5. 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.

sanzaru CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
sanzaru CLAUDE.mdTJC-LP/sanzaru	Instructions file	6	yesterday	MIT	A	19,838	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	yesterday	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

sanzaru: Instructions file for Claude Code