DOCUMENTATION文档
Run Sieve in five minutes. 五分钟跑起 Sieve。
Sieve is a local LLM-traffic security proxy — one Rust binary between your coding agent and the LLM API. These docs cover install, the detection engine, and how to verify every claim yourself. Sieve 是一个本地 LLM 流量安全代理——夹在你的编码 agent 与大模型 API 之间的一个 Rust 二进制。这份文档涵盖安装、检测引擎，以及如何亲手验证每一条声明。
Quickstart快速开始
$ brew install sieve # or: curl -fsSL sieveai.dev/install.sh | bash
$ sieve setup
$ sieve doctor && sieve decisions watch
INSTALL
Install & verify安装与验证
One-line install that cosign-verifies the binary for you (fail-closed), plus how to re-verify by hand.
一行安装，安装器替你 cosign 验签（fail-closed），以及如何亲手复验。
CONCEPT
Outbound redaction出站脱敏
How prompts are scanned and secrets are replaced in place before they leave 127.0.0.1.
提示词如何被扫描，密钥如何在离开 127.0.0.1 之前就地替换。
CONCEPT
Inbound interception入站拦截
How responses are held, and how Critical tool calls fail closed without a human beat.
响应如何被暂留，Critical 工具调用如何在无人确认时 fail-closed。
RULES
The rule engine规则引擎
Severity levels, the built-in crypto ruleset, and the format for writing your own.
严重级别、内置加密规则集，以及编写自有规则的格式。
CLI
Command reference命令参考
sieve setup, decisions watch, rules test, and every flag with examples.
sieve setup、decisions watch、rules test，以及每个参数的示例。
VERIFY
Reproduce the numbers复现这些数字
Run the public benchmark suite and reproduce the 0.00% / 99.71% figures locally.
运行公开基准套件，在本地复现 0.00% / 99.71% 的数据。