AI Skill 大全

终端命令与代码仓库安全检查

输入待执行命令、网址或代码仓库,识别恶意下载、仿冒域名和隐藏字符,输出风险判断及触发原因。

功能与用途

用于执行陌生命令、下载脚本或接手外部代码前的安全检查。提供命令、网址、文件或仓库路径,即可检查仿冒域名、隐藏字符、终端控制符和配置投毒,获得风险等级、命中规则及检查报告。还可检查本地 Python 包文件、追查拦截原因,并导出审计记录。需要先安装 tirith;持续拦截需配置受支持的终端或 AI 编程客户端,并验证防护已生效。仅注册 MCP 服务不代表自动拦截。

适用场景

  • 执行安装命令前排查风险:检查 curl 管道执行脚本等命令,识别危险执行模式、仿冒下载地址和不安全传输。
  • 接手外部仓库前扫描:扫描仓库中的不可见字符、终端控制符和配置投毒,生成便于审查的风险报告。
  • 为 AI 编程客户端接入防护:配置客户端支持的执行钩子或 MCP 网关,并验证命令检查与拦截是否生效。
  • 调查命令被拦截的原因:查看最近触发的规则和审计记录,核对误报后设置限定规则与有效期的信任项。

适用人群

  • 经常运行外部安装脚本的开发者
  • 使用 AI 编程代理的工程师
  • 负责代码安全审查的安全工程师
  • 维护开发环境与流水线的运维工程师

使用步骤

  1. 提供待检查命令、网址或本地路径。
  2. 选择命令检查、网址评分或文件扫描。
  3. 查看风险等级、命中规则和触发原因。
  4. 按需导出报告,或配置持续防护并验证生效。

开始使用

将需求和 Skill 名称发送给已连接 SkillPX MCP 的 AI 助手,核对适用工具和安装要求。

查看 MCP 接入说明

技术信息

分类
IT运维与安全
适用工具
Codex、Claude Code、OpenClaw、Hermes Agent、国内 Agent
使用难度
进阶
版本
main
最近更新
2026-11-03
文件数
81
安装包大小
自动生成 ZIP
安全检查
基础检查通过

>

使用前准备

  • 已安装 tirith,且当前终端可以调用其命令。
  • 准备待检查的命令、网址、本地文件、包文件或仓库路径。
  • 启用持续防护时,需要配置受支持的终端钩子或 AI 客户端集成,并验证实际生效。
  • 脚本下载分析与运行功能限 Unix;不可信项目隔离运行限受支持的 x86_64 Linux。

依赖与权限

  • [workspace]
  • members = [
  • "crates/tirith-core",
  • "crates/tirith",
  • "crates/tirith-test-support",
  • "tools/sign-license",
  • "tools/license-server",
  • "tools/policy-server",
  • ]
  • resolver = "2"
  • [workspace.package]
  • version = "0.5.0"
  • edition = "2021"
  • license = "AGPL-3.0-only"
  • rust-version = "1.83"
  • [workspace.dependencies]
  • tirith-core = { version = "0.5.0", path = "crates/tirith-core" }
  • tirith-test-support = { path = "crates/tirith-test-support" }
  • clap = { version = "4", features = ["derive"] }
  • url = "2"
  • regex = "1"
  • regex-syntax = "0.8"
  • percent-encoding = "2"
  • serde = { version = "1", features = ["derive"] }
  • serde_json = "1"
  • serde_yaml = "0.9"
  • toml_edit = "=0.22.27"
  • once_cell = "1"
  • base64 = "0.22"
  • etcetera = "0.8"
  • [package]
  • name = "tirith"
  • version.workspace = true
  • edition.workspace = true
  • license.workspace = true
  • rust-version.workspace = true
  • description = "Terminal security - catches homograph attacks, pipe-to-shell, ANSI injection"
  • repository = "https://github.com/sheeki03/tirith"
  • homepage = "https://github.com/sheeki03/tirith"
  • readme = "../../README.md"
  • keywords = ["security", "url", "shell", "cli"]
  • categories = ["command-line-utilities"]
  • [[bin]]
  • path = "src/main.rs"
  • name = "tirith-threatdb-compile"
  • path = "src/bin/tirith_threatdb_compile.rs"
  • name = "tirith-package-approval-authority"
  • path = "src/bin/tirith_package_approval_authority.rs"
  • [features]
  • default = []
  • artifact-hash-lookup = ["tirith-core/artifact-hash-lookup"]
  • sigstore-attestations = []
  • [dependencies]
  • tirith-core = { workspace = true }
  • clap = { workspace = true }
  • clap_complete = { workspace = true }
  • clap_mangen = { workspace = true }
  • name = "tirith-core"
  • description = "Terminal security analysis engine - homograph attacks, pipe-to-shell, ANSI injection"
  • keywords = ["security", "url", "analysis"]
  • categories = ["development-tools"]
  • artifact-hash-lookup = []
  • test-network-seams = []
  • url = { workspace = true }
  • regex = { workspace = true }
  • regex-syntax = { workspace = true }
  • percent-encoding = { workspace = true }
  • serde = { workspace = true }
  • serde_json = { workspace = true }
  • serde_yaml = { workspace = true }
  • once_cell = { workspace = true }
  • base64 = { workspace = true }
  • etcetera = { workspace = true }
  • home = { workspace = true }
  • chrono = { workspace = true }
  • sha2 = { workspace = true }
  • name = "tirith-fuzz"
  • version = "0.0.0"
  • publish = false
  • [package.metadata]
  • cargo-fuzz = true
  • libfuzzer-sys = "0.4"
  • tirith-core = { path = "../crates/tirith-core" }
  • members = ["."]
  • name = "url_parser"
  • path = "fuzz_targets/url_parser.rs"
  • doc = false
  • name = "tokenizer_posix"
  • path = "fuzz_targets/tokenizer_posix.rs"
  • name = "tokenizer_fish"
  • path = "fuzz_targets/tokenizer_fish.rs"
  • name = "tokenizer_powershell"
  • path = "fuzz_targets/tokenizer_powershell.rs"

同类 Skill