Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问许可证需确认审计异常

stack-audit堆栈审计

Agent Skill

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

总安装

212

周安装

9

GitHub Stars

26

下载量

74
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:stack-audit(堆栈审计)
来源仓库:https://github.com/outfitter-dev/agents
仓库路径:skills/stack-audit
安装命令:
npx skills add https://github.com/outfitter-dev/agents --skill stack-audit
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/outfitter-dev/agents --skill stack-audit

简介

用于辅助安全审计与权限风险排查。stack-audit 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

  • 可梳理敏感配置、检查依赖风险或分析鉴权逻辑。
  • 不能将工具输出直接当作最终结论。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。
  • 涉及密钥或生产系统时应确认最小权限与脱敏方式。
  • 不建议直接修改认证流程或访问未授权数据。

SKILL.md

Stack Audit

Scan a codebase to identify Outfitter Stack adoption candidates and generate an audit report.

Quick Start

Option 1: Run the scanner (recommended for large projects)

bun run plugins/outfitter-stack/skills/stack-audit/scripts/init-audit.ts [project-root]

Generates .outfitter/adopt/ with:

  • audit-report.md - Scan results and scope
  • plan/ - Stage-by-stage task files

Option 2: Manual scan (smaller projects)

Run the audit commands below to understand scope.

Audit Commands

Critical Issues - Exceptions

# Count throw statements
rg "throw (new |[a-zA-Z])" --type ts -c

# List throw locations
rg "throw (new |[a-zA-Z])" --type ts -n

# Count try-catch blocks
rg "(try \{|catch \()" --type ts -c

Console Usage

# Count console statements
rg "console\.(log|error|warn|debug|info)" --type ts -c

# List console locations
rg "console\.(log|error|warn|debug|info)" --type ts -n

Hardcoded Paths

# Homedir usage
rg "(homedir\(\)|os\.homedir)" --type ts -c

# Tilde paths
rg "~/\." --type ts -c

# Combined path issues
rg "(homedir|~\/\.)" --type ts -n

Custom Error Classes

# Find custom error classes
rg "class \w+Error extends Error" --type ts -n

# Count usage of custom errors
rg "new MyCustomError\(" --type ts -c

Generated Structure

.outfitter/adopt/
├── audit-report.md           # Scan results, scope, recommendations
└── plan/
    ├── 00-overview.md        # Status dashboard, dependencies
    ├── 01-foundation.md      # Dependencies, context, logger
    ├── 02-handlers.md        # Handler conversions
    ├── 03-errors.md          # Error taxonomy mappings
    ├── 04-paths.md           # XDG path migrations
    ├── 05-adapters.md        # CLI/MCP transport layers
    ├── 06-documents.md       # Documentation updates
    └── 99-unknowns.md        # Items requiring review

Migration Stages

StageBlocked ByFocus
1. Foundation-Install packages, create context/logger
2. HandlersFoundationConvert throw to Result
3. ErrorsHandlersMap to error taxonomy
4. Paths-XDG paths, securePath
5. AdaptersHandlersCLI/MCP wrappers
6. DocumentsAllUpdate docs to reflect patterns
99. Unknowns-Review anytime

Audit Report Fields

FieldDescription
Exceptionsthrow statements to convert to Result
Try/CatchError handling blocks to restructure
ConsoleLogging to convert to structured logging
PathsHardcoded paths to convert to XDG
Error ClassesCustom errors to map to taxonomy
HandlersFunctions with throws to convert
UnknownsComplex patterns requiring review

Error Taxonomy Reference

When mapping errors, use this reference:

OriginalOutfitterCategory
NotFoundErrorNotFoundErrornot_found
InvalidInputErrorValidationErrorvalidation
DuplicateErrorConflictErrorconflict
UnauthorizedErrorAuthErrorauth
ForbiddenErrorPermissionErrorpermission
Generic ErrorInternalErrorinternal

Effort Estimation

CountEffort Level
0None
1-5Low
6-15Medium
16+High

Interpreting Results

High-Priority Items

  • Functions with 3+ throw statements (complex error handling)
  • Files with 3+ try-catch blocks (may need restructuring)
  • Custom error classes with high usage counts

Medium-Priority Items

  • Isolated throw statements (simple conversions)
  • Console logging (straightforward migration)
  • Hardcoded paths (mechanical replacement)

Low-Priority Items

  • Documentation updates (can happen last)
  • Test file updates (follow handler changes)

Next Steps After Audit

  1. Review audit-report.md for accuracy
  2. Adjust priorities in plan/00-overview.md
  3. Begin with Stage 1 (Foundation)
  4. Load outfitter-stack:stack-patterns for conversion guidance
  5. Load outfitter-stack:stack-templates for scaffolding

Constraints

Always:

  • Run audit before planning adoption
  • Review unknowns for complex patterns
  • Estimate effort before committing

Never:

  • Skip the audit phase
  • Underestimate try-catch complexity
  • Ignore custom error classes

Related Skills

  • outfitter-stack:stack-patterns - Target patterns reference
  • outfitter-stack:stack-templates - Component templates
  • outfitter-stack:stack-review - Verify compliance

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

36.24%
按下载量换算27

Claude

31.1%
按下载量换算23

Cursor

19%
按下载量换算14

Gemini CLI

10.41%
按下载量换算8

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills