What is VLM-mcp AGENTS.md?

A set of project instructions for VLM-mcp, a local server that lets an agent work with image-understanding models. It documents the project setup, important files, configuration, caching, sessions, and commands.

Use it when working on the VLM-mcp Python project, especially its server, model launcher, configuration, sessions, caches, and image handling.

It gives an agent the project-specific rules needed to navigate and modify the code consistently.

How YC-CLT/VLM-mcp configures Codex and OpenCode

Reuse

Borrowing it

Nothing to install: this file belongs to YC-CLT/VLM-mcp. 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/YC-CLT/VLM-mcp/main/AGENTS.md

Clone the repo

git clone --depth 1 https://github.com/YC-CLT/VLM-mcp

Made for: Codex, OpenCode.

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/yc-clt/vlm-mcp/agents-md/github.svg)](https://agentmods.dev/instructions/yc-clt/vlm-mcp/agents-md)

Your own site

<a href="https://agentmods.dev/instructions/yc-clt/vlm-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/yc-clt/vlm-mcp/agents-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/yc-clt/vlm-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/yc-clt/vlm-mcp/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 2,002 This file is loaded in full into every session.

When invoked 2,002 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	$0.02002	$0.02002
Opus 5.5	$0.00801	$0.00801
Sonnet 5.5	$0.00400	$0.00400
Haiku 4.5	$0.00200	$0.00200

Measured 1mo ago against content hash a75ed2776258, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

VLM-mcp AGENTS.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.

AGENTS.md · 90 lines

How it starts

The opening of the file, as written. The whole thing — 90 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

环境

Python 3.11，uv 管理依赖
config.py 集中配置所有常量，pyproject.toml 管理依赖
命令：用 cmd-exec-mcp

关键文件

文件	作用
`main.py`	入口，纯启动脚本，启动 MCP
`server.py`	MCP Server，7 个工具，stdio 传输
`config.py`	所有常量 + `config.json` 加载
`config.json`	敏感配置（API Key、模型路径），不入 git
`llama_launcher.py`	llama-server 子进程生命周期（启/停/健康检查）
`providers/ocr_provider.py`	OCR Provider（RapidOCR 懒加载 + 延迟卸载）
`providers/openai_compat.py`	OpenAI 兼容 API 客户端（AsyncOpenAI）+ llama 懒启动
`session_manager.py`	会话 CRUD + 超时清理 + 满时驱逐
`cache.py`	L1 图片缓存（LRU）+ L2 响应缓存（LRU + TTL）
`image_utils.py`	图片源解析（路径/URL/Data URI/Base64 回退）
`logger.py`	单例 logger → `log.txt`

关键常量

常量	位置	说明
`IMAGE_MAX_SIZE_MB`	`config.py`	图片最大体积 (20MB)
`IMAGE_DOWNLOAD_TIMEOUT`	`config.py`	图片下载超时秒数 (10s)
`CACHE_IMAGE_MAX_ENTRIES`	`config.py`	L1 缓存上限 (100)
`CACHE_RESPONSE_MAX_ENTRIES`	`config.py`	L2 缓存上限 (500)
`CACHE_RESPONSE_TTL_ONLINE`	`config.py`	在线后端缓存 TTL (3600s)
`CACHE_RESPONSE_TTL_LOCAL`	`config.py`	本地后端缓存 TTL (1800s)
`SESSION_TTL`	`config.py`	会话超时秒数 (1800s)
`SESSION_MAX`	`config.py`	每后端最大会话数 (5)
`LOG_LEVEL`	`config.py`	日志级别 ("INFO")
`LLAMA_DEFAULTS`	`config.py`	llama-server 启动参数默认值
`BACKENDS`	`config.py`	从 `config.json` 加载的后端配置
`CACHE_ENABLED`	`config.py`	从 `config.json` 加载的缓存开关

规则

monkeypatch 必须用 import config + config.X：from config import X 创建本地副本，monkeypatch 无法穿透；executor 同理 patch executors.模块名.X
config 重命名全量 grep：常量改名/移除后搜索所有引用
跨 Task 依赖等待：并行派发时先检查上游产物是否存在

工具

部分工具有相应的skill
MCP详见 mcp_tools_summary.csv
Read 无法访问 D:\Temp，MCP 长输出需 Copy-Item 到项目根目录，正则替换 \\n 为 \n
Edit replace_all 错误：replace_all=True 无法使用。优先用 PowerShell Select-String + 正则做精确替换，或手动逐处 Edit
浏览器操控用 chrome-devtools-edge（Edge CDP），禁止用 cua-driver 操控浏览器
WebFetch 无法使用：用 wet-mcp extract

工作流

Read the full file on GitHub · 90 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 · +3 lines · +176 tokens per session a75ed2776258
2mo ago Changed · +5 lines · +226 tokens per session 21058e7157da
2mo ago First seen · 82 lines · 1,600 tokens per session scan A 4a427c014cff

Subscribe to this mod's changes

VLM-mcp AGENTS.md is an instructions file published in the GitHub repository YC-CLT/VLM-mcp (0 stars, last pushed 1mo ago), licensed MIT. It adds 2,002 tokens to every session, about $0.0080 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.

VLM-mcp AGENTS.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
VLM-mcp AGENTS.mdYC-CLT/VLM-mcp	Instructions file	0	1mo ago	MIT	A	2,002	Codex, OpenCode
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	9,719	GitHub Copilot
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	5,516	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	today	MIT	A	4,680	Codex, OpenCode
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	3,093	GitHub Copilot
vscode model-prompts.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	2,022	GitHub Copilot

VLM-mcp: Instructions file for Codex