sofagent GEMINI.md

A GEMINI.md instruction file for Sofagent, an adapter that connects the Sofagent rules and tools to Gemini CLI. It defines the order for loading instructions, knowledge, audits, and MCP tools.

In plain words
What is it for?
Configuring Sofagent for Gemini CLI, loading its layered instructions, running commit audits, and connecting to the Sofagent MCP server.
Why use it?
It ensures Gemini CLI sessions use the intended project rules and required checks rather than relying on incomplete or duplicated guidance.

Instructions file for Gemini CLI

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/kongfangxun/sofagent/gemini-md
Clone the repo
git clone --depth 1 https://github.com/KongFangXun/sofagent

Made for: Gemini CLI.

Per session 464 This file is loaded in full into every session.
When invoked 464 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 $0.00464 $0.00464
Opus 5 $0.00232 $0.00232
Sonnet 5 $0.00093 $0.00093
Haiku 4.5 $0.00046 $0.00046

Measured 2d ago against content hash 334db1a8fa06, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

sofagent GEMINI.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 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.

GEMINI.md · 34 lines

What it actually says

GEMINI.md · sofagent Gemini CLI 适配(薄挂载)

本文件是 Gemini CLI 的分层上下文挂载点(GEMINI.md 约定),不含规范内容——四层加载链的规范资产唯一来源是 sofagent 的 SKILL/ 目录(单一真相源)。规则更新只发生在 SKILL/,本文件永不复制它们。

会话开始时按序加载(四层加载链)

一、读取已安装的 SKILL.md 作为系统提示词(L1 硬约束——Gemini CLI 将本文件与引用内容一并注入系统上下文):

~/.sofagent/skills/sofagent/SKILL.md

(若用 bash install.sh --platform gemini 安装,则为 ~/.gemini/skills/sofagent/SKILL.md。)

二、读取用户工作规则(L2):~/.sofagent/skills/sofagent/fde.md

三、读取反思区(L3):~/.sofagent/skills/sofagent/think.md

四、知识库(L4):~/.sofagent/skills/sofagent/knowledge/ 目录,按任务关键词检索加载。

审计强制(平台无关)

约束是建议性的,审计是强制性的——提交前审计走 git hooksofagent-audit --install-hook),与宿主平台无关:24 条 git diff 规则 + HMAC 链审计在 Gemini CLI 下同样生效。

连接 MCP Server

Gemini CLI 的 MCP 配置(settings.json)指向 sofagent MCP Server(stdio):

{ "mcpServers": { "sofagent": { "command": "sofagent-mcp" } } }

61 个 tool(run_audit / worklog_query / snapshot_restore 等)经 MCP 协议面暴露——适配器只依赖 @public API 子集,@internal 破坏性变更不影响本挂载。

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 · 34 lines · 464 tokens per session scan A 334db1a8fa06

Subscribe to this mod's changes

sofagent GEMINI.md is an instructions file published in the GitHub repository KongFangXun/sofagent (41 stars, last pushed 4d ago), licensed MIT. It adds 464 tokens to every session, about $0.0023 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.