fcop-protocol

A collaboration protocol that defines how AI agents record work, organize files, and move tasks through stages.

In plain words
What is it for?
It is for coordinating agent tasks, applying naming and file-format rules, and managing a shared task lifecycle.
Why use it?
It reduces confusion between agents by making decisions and progress visible in files instead of leaving them only in conversation.

Cursor rule for Cursor

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 rules/joinwell52-ai/codeflowmu-open/fcop-protocol
Clone the repo
git clone --depth 1 https://github.com/joinwell52-AI/CodeFlowMu-open

Made for: Cursor.

Per session 34,375 This file is loaded in full into every session.
When invoked 34,375 The same file — it is already loaded in full.
Security scan E 2 findings. Scan, not verified.
Origin 97% copy Near-identical to another mod 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 $0.34375 $0.34375
Opus 5 $0.17188 $0.17188
Sonnet 5 $0.06875 $0.06875
Haiku 4.5 $0.03438 $0.03438

Measured 2d ago against content hash 07ec39a0d524, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade E, and why

fcop-protocol scanned grade E with 2 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.

Hidden instructionshighPrompt injection

Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.

<!-- Host-neutral reminder / 宿主中立提示: The conventions below describe FCoP itself (a file-based coordination protocol), not anything specific to Cursor. The `.mdc` wrapper and `alwaysApply` frontmatter are a Cursor-specifi

Reaches for credential fileshighPrivilege escalation

SSH keys, cloud credentials, git-credentials, .npmrc, /etc/shadow: reading these is how a config file becomes a credential leak.

- PM 在 TASK 文档里写 G6 描述:"cached diff 对 `\.env|\.aws/credentials` 0 命中"
Origin

This is a copy

97% identical to fcop-protocol — 20 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.

templates/default-project/.cursor/rules/fcop-protocol.mdc · 2,357 lines

How it starts

The opening of the file, as written. The whole thing — 2,357 lines — stays where its author put it; the contents beside it link to each section on GitHub.

FCoP Protocol · 协议解释 / Protocol Commentary

本文件是 FCoP 的协议解释——把 fcop-rules.mdc 里那 10 条协议规则 (Rule 0–9)落到实际场景里:文件怎么命名、YAML 怎么写、目录怎么组织、 巡检怎么触发。两个文件同为 alwaysApply: true,但当本文件与 协议规则冲突时,fcop-rules.mdc 为准

This file is the protocol commentary on fcop-rules.mdc — how each rule (0–9) actually applies in practice: file naming, YAML shape, directory layout, patrol triggers. Both files are alwaysApply: true. In case of conflict, fcop-rules.mdc wins.

v1.0 final(fcop_protocol_version 1.9.0):Rule 9 的完整 commentary 已补入文末"Rule 9 Commentary"节。七大核心概念的规范性定义 见 spec/archived/fcop-runtime-protocol-v1.0.md(英文)/ .zh.md(中文)。

v1.0 final (fcop_protocol_version 1.9.0): Full Rule 9 commentary is now in the "Rule 9 Commentary" section at the end of this file. Normative definitions of the seven core concepts: spec/archived/fcop-runtime-protocol-v1.0.md (EN) / .zh.md (ZH).

Core Principle / 核心原则

AI agents must not talk only inside their heads — they must land it as a file.

AI 角色之间不能只在脑子里说话,必须落成文件。

This is the overall principle of the entire protocol. Every specific rule below — file naming, YAML frontmatter, subtask batches, collaboration rules — can be understood as "this principle landing in a specific scenario". The principle itself was not designed top-down; it was surfaced by an agent during an unrelated task and then adopted in reverse as the overall rule.

本条为整份协议的总则。下面所有的具体规定——文件命名、YAML 元数据、分包任务、 协作规则——都可以被理解为"这条原则在不同场景下的具体落地"。这条总则不是自上 而下设计出来的,是某次无关任务中由 agent 自发升华得出、后被反向收回为总则。

Read the full file on GitHub · 2,357 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.

  1. 2d ago First seen · 2,357 lines · 34,375 tokens per session scan E 07ec39a0d524

Subscribe to this mod's changes

fcop-protocol is a cursor rule published in the GitHub repository joinwell52-AI/CodeFlowMu-open (2 stars, last pushed 8d ago), licensed MIT. It adds 34,375 tokens to every session, about $0.1719 per session on Opus 5. A static security scan graded it E with 2 findings (hidden instructions, reaches for credential files). It is 97% identical to fcop-protocol, differing in 20 lines, and is treated as a copy.