agent-claim-network AGENTS.md

agent-claim-network AGENTS.md is an instructions file for Codex, OpenCode from FTShare-Lab/agent-claim-network. It costs 1,781 tokens per session, scanned A, original, Apache-2.0.

A repository guide for AI coding agents working on Agent Claim Network, a Rust terminal application for sharing claims, policies, disputes, and team knowledge between assistants.

In plain words
What is it for?
Guiding work on the local agent, API, router, maintainer, claim data, configuration, and terminal interface.
Why use it?
It explains the system's parts, data boundaries, documentation sources, and Rust conventions so agents can make changes in the right place.

Instructions file for CodexOpenCode

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 instructions/ftshare-lab/agent-claim-network/agents-md
Clone the repo
git clone --depth 1 https://github.com/FTShare-Lab/agent-claim-network

Made for: Codex, OpenCode.

Per session 1,781 This file is loaded in full into every session.
When invoked 1,781 The same file — it is already loaded in full.
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 $0.01781 $0.01781
Opus 5 $0.00890 $0.00890
Sonnet 5 $0.00356 $0.00356
Haiku 4.5 $0.00178 $0.00178

Measured 3d ago against content hash 8af1a82b822f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-claim-network 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 3d 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 · 87 lines

How it starts

The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

本文档为 Codex 在本仓库中工作时提供指引。整体使用中文与用户交互。

项目概览

Agent Claim Network(ACN)是 Rust 实现的通用领域协作型助手,主要通过终端 TUI 与用户交互。acn 可独立运行,也可连接 Router 与 Maintainer,共享 claim、接收 policy、上报 dispute 并维护团队知识。

以当前代码与现行文档为准,不以历史设计稿或已完成 PRD 的实施清单为准。按任务查阅:

  • README.md / docs/user_guide.md
  • docs/architecture.md
  • docs/core_behavior.md
  • docs/memory_design.md
  • docs/config_parameters.md
  • docs/PRDs/

需求或行为有歧义时先向用户确认,不替业务方拍板。

当前架构

模块细节见 docs/architecture.mdsrc/;改动时守住所有权即可:

  • Agent(src/agent/src/session/src/memory.rssrc/session_tui/src/tool/):本地 session、Memory、TUI 与工具执行;inbox 内化后才落本地变更。
  • API(src/api/):provider-neutral turn loop 与各模型 adapter。
  • Router(src/router/):可引用的 claim / dispute 检索与派生视图,不拥有 agent 私有 Memory。
  • Maintainer(src/maintainer/):policy、dispute、stale sweep、outbox;对 claim 属性的更新是建议而非强制。
  • Claim DTO(src/claim/)与装配入口(src/bin/acn.rssrc/bootstrap.rs)。

默认 ACN base 目录是 ~/.acn。upstream 是 Agent 侧的运行与团队连接配置;Agent 选中后,私有运行时根目录是 <acn_home>/<upstream>/。Router 与 Maintainer 不解析或选择 Agent upstream,团队数据固定使用各自 <acn_home>/data/team/。工作区由 --cd 或启动 cwd 决定,不能与 ACN 数据目录混淆。

Rust 规范

  • 异步 I/O 使用 tokio;禁止在 async 上下文直接做阻塞网络、文件、子进程或 sleep。
  • async 锁优先 tokio::sync::Mutex,不要持有同步锁跨越 .await
  • 应用入口使用 anyhow::Result;模块边界优先 thiserror
  • 避免 unwrapexpectpanic!todo!unimplemented!。确有静态安全前提时就近写清原因。
  • 降位转换使用 TryFrom / TryInto,不要直接 as
  • 路径使用 Path / PathBufjoin,不要拼接路径字符串。
  • YAML 使用 serde-yaml-ng。持久状态写入遵循现有原子写、文件锁和恢复边界,不自行发明第二套。
  • 复用现有配置结构;重要 endpoint、路径、模型和超时不在业务代码硬编码。
  • 日志沿用项目现有 log/ftlog 管线,按可恢复程度选择 error、warn、info、debug、trace。
  • 文件过大或职责混杂时拆到现有同级模块,避免为单文件创建无意义层级。
  • 模块顶部用简短中文说明职责;注释解释约束和原因,不复述代码。

数据与安全边界

  • USER.md 永不进入 claim;MEMORY.md 是 agent 私有内容,不直接上传团队服务。
  • Router 只返回可引用的候选 claim 与 dispute 信息,不拥有 agent 私有 Memory。
  • Maintainer 的 claim 属性更新是建议;agent 在 inbox 内化后决定本地变更。
  • 单人模式不得发起团队网络请求,也不为未来补传偷偷积累团队上传队列。
  • Team Auth、MCP token、LLM key 只保存环境变量名;仓库任何位置都不得出现明文密钥或其他真实 secret,包括测试、fixture、示例、文档、日志和错误信息。
  • 仓库任何位置都不得出现内网域名;localhost127.0.0.1 等明确的本机回环地址除外。测试与示例需要域名时使用标准保留域名或中性占位符。
  • 路径、变量名、fixture 和示例数据不得包含可能暴露真实用户名、姓名缩写或内部账号的信息;使用 userexample、临时目录等中性名称。
  • Rust、JavaScript/TypeScript、Python、Shell等代码文件中,不以 Codex/GPT/OpenAI、Claude/Claude Code/CC/Anthropic、Hermes/HA/Hermes Agent、GA/Generic Agent、非凸/非凸科技/ft.tech/non-convex/ftai 作为实现来源、兼容对象、架构标签、变量/fixture 名或注释描述。运行时真实默认值及协议、模型、API、序列化标识等兼容性所必需的名称除外,例如 gpt-5.5openai_chat_compatible
  • Markdown 与 HTML 文档不受上一条外部主体名称限制,但仍必须遵守明文密钥、内网域名和真实用户信息的禁令。
  • 修改协议 DTO 前检查持久化兼容、HTTP 客户端/服务端、prompt 与相关测试。

Read the full file on GitHub · 87 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. 3d ago First seen · 87 lines · 1,781 tokens per session scan A 8af1a82b822f

Subscribe to this mod's changes

agent-claim-network AGENTS.md is an instructions file published in the GitHub repository FTShare-Lab/agent-claim-network (41 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,781 tokens to every session, about $0.0089 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.