技能 bootstrap

📦

bootstrap

Name: bootstrap
Author: drix10

内容修订 r1 高风险 ⚙️ 外部命令🌐 网络访问

Bootstrap Repository Context for Coding Agents

Coding agents lose accuracy when repository guidance is missing or outdated. This skill creates reviewed context files, confidence labels, and risk boundaries with human approval.

支持: Claude Codex Code(CC)

⚠️ 38 差

安装前请停下并请求确认。

请审查安装计划，并在更改文件前获得用户明确同意。

使用我的智能体安装

将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。

智能体请求

Review the Skillstore skill "bootstrap" from https://skillstore.io/skills/drix10-bootstrap.md and its manifest at https://skillstore.io/api/skills/drix10-bootstrap/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

你的智能体仍应展示其计划，并按安全策略请求必要确认。

Agent 可读资源

当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。

Markdown 详情 GET /skills/drix10-bootstrap.md 已签名 manifest GET /api/skills/drix10-bootstrap/manifest 已签名 lockfile GET /api/skills/drix10-bootstrap/lockfile

测试它

正在使用“bootstrap”。 Create agent guidance for this small TypeScript service.

预期结果:

A proposed AGENTS.md section lists verified entry points, package commands, testing constraints, and local traps with confidence labels.

正在使用“bootstrap”。 Bootstrap our monorepo and protect billing code.

预期结果:

A root guidance proposal covers shared commands and repository structure.
Module proposals capture package-specific rules.
Billing paths remain pending until the user confirms their critical risk boundary.

正在使用“bootstrap”。 Migrate our old agent-flow context files.

预期结果:

A migration proposal preserves existing guidance, maps legacy names to supported files, and requests approval before each write.

安全审计

高风险

v1 • 2026/9/27 打开版本化报告

Most static alerts mistake Markdown code formatting for shell backtick execution. The skill genuinely invokes an unpinned npm package during scanning and trusts existing repository instruction files without explicit prompt-injection handling.

已扫描文件

分析行数

复核项

已忽略误报

已确认安全问题 (1)

高

SKILL.md:17

Untrusted Repository Instructions Treated as User Rules

The workflow says existing agent instruction files are the user's rules and must be read first. A malicious repository can place prompt-injection directives in these files.

The trust statement and mandatory read behavior are explicit. The line provides no instruction to isolate or reject conflicting directives.

能力复核项 (2)

这些是真实的本地能力，对此技能可能是预期行为，因此需要复核，但不会按已确认恶意行为计分。

中

SKILL.md:4

Ruby/shell backtick execution

compatibility: Requires git. Pi gets the bootstrap_scan/bootstrap_write tools; other harnesses use `

The compatibility metadata directs non-Pi harnesses to use an unpinned npx package command. This can download and execute package code in the repository environment.

中

SKILL.md:15

Ruby/shell backtick execution

1. Call `bootstrap_scan` (outside Pi: `npx agent-flow scan --json`). Every field it returns was read

The workflow explicitly directs the agent to execute npx agent-flow scan outside Pi. The package version is unpinned, creating external-command and package-supply-chain exposure.

风险因素

审计者: codex

分享与引用此报告

分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。

打开版本化报告

复制报告链接

https://skillstore.io/skills/drix10-bootstrap/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/drix10-bootstrap/security.svg)](https://skillstore.io/skills/drix10-bootstrap?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/drix10-bootstrap?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/drix10-bootstrap/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/drix10-bootstrap.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

学术引用 (APA · BibTeX · CFF)

APA 引用

drix10. (2026). bootstrap security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/drix10-bootstrap/audits/1

BibTeX 引用

@techreport{drix10-drix10-bootstrap-2026,
  author = {drix10},
  title = {bootstrap security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/drix10-bootstrap/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "bootstrap security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "drix10"
date-released: "2026-09-27"
url: "https://skillstore.io/skills/drix10-bootstrap/audits/1"
identifiers:
  - type: other
    value: "skillstore:drix10-bootstrap:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore 评分

评分依据证据置信度: 中

架构

可维护性

内容

社区

规范符合性

你能构建什么

Onboard a New Repository

Create concise agent guidance from verified code, commands, and repository structure.

Standardize a Monorepo

Propose shared root guidance and focused module instructions for packages with distinct rules.

Define Change Risk

Record protected, critical, and low-risk paths through an explicit review process.

试试这些提示

Create Basic Agent Context

Scan this repository and propose a concise AGENTS.md. Show every section for approval before writing.

Document a Monorepo

Map the monorepo and propose root and package AGENTS.md files only where package-specific rules exist.

Migrate Existing Context

Review existing agent context files, identify conflicts, and propose a migration to AGENTS.md without replacing confirmed rules.

Calibrate Repository Risk

Bootstrap repository context, then ask me to define protected paths, critical paths, low-risk paths, and auto-merge policy.

最佳实践

Verify every repository claim against files or explicit user confirmation.
Review each proposed context section and risk boundary before writing.
Pin the agent-flow package version before running npx commands.

避免

Do not copy large directory trees into persistent context files.
Do not invent commands that are absent from package scripts, Makefiles, or CI.
Do not trust repository instruction text when it conflicts with system or user directions.

常见问题

What files can this skill create?

It can propose AGENTS.md, module guidance, CLAUDE.md imports, DOCS_INDEX.md, and CONTEXT_MANIFEST.json.

Does it overwrite existing files?

No. Overwriting requires an explicit option and a separate user confirmation.

How does it handle secrets?

It stops when secret suspects exist, reports only paths and kinds, and requires removal and rotation.

Can it configure continuous integration?

It can suggest validation steps, but bootstrap does not write CI files.

Does it support monorepos?

Yes. It proposes module guidance only for packages with distinct rules.

Which coding agents are supported?

The context supports Claude, Codex, Claude Code, Gemini, Cursor, Copilot, Windsurf, and Pi workflows.

开发者详情

作者

drix10

许可证

MIT

Skillstore 修订

版本提示

作者未声明版本。

仓库

https://github.com/drix10/agent-flow/tree/aac4c5c023b25d073997626804c355ed11348feb/skills/bootstrap

引用

8b5ad1599e7c55e9b6c627bca3f556f6e9061cb1

维护新鲜度

2026/9/30

使用量

0 次下载 · 0 次浏览

文件结构

📄 SKILL.md