Token导航 LogoToken导航

Agent Skill · 研究检索

bitbucket-toolbox(位桶工具箱)

bitbucket-toolbox 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

bitbucket-toolbox 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 20,970。

总安装 5,018周安装 207GitHub Stars 公开资料未说明下载 1,639核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
1.1.2

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:bitbucket-toolbox(位桶工具箱)
来源仓库:https://github.com/zan768616253/bitbucket-toolbox
安装命令:
openclaw skills install bitbucket-toolbox
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install bitbucket-toolbox

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于安全审查 Bitbucket Cloud 的拉取请求,支持按文件拆分差异和代码分析。适合在 OpenClaw 中需要根据关键词快速定位候选结果或筛选相关信息时使用。通过 API 包装器优化代码审查流程,辅助代理完成 Pull Request 审阅任务。

安装命令:openclaw skills install bitbucket-toolbox,需确认权限范围和 API 密钥配置。注意检查是否会触发网络请求或文件操作,避免越权访问敏感仓库内容。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install bitbucket-toolbox;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

bitbucket-toolbox 怎么安装?

在终端执行 openclaw skills install bitbucket-toolbox。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

bitbucket-toolbox 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

bitbucket-toolbox 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

bitbucket-toolbox 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

bitbucket-toolbox 用的人多吗?

公开数据是累计安装 5,018 次,近一周安装 207 次,下载 1,639 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

bitbucket-toolbox 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
bitbucket-toolbox

位桶工具箱 · 本页

5,018—接触敏感数据MIT-0通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

name
bitbucket-toolbox
version
1.1.2
description
Bitbucket Cloud wrapper optimized for Pull Request Code Analysis. Enables the agent to securely review Pull Requests, split large diffs by file, review code structure, and read specific repository files. Ideal for providing automated code reviews or debugging PRs.
author
Eric Wang
license
MIT
homepage
https://github.com/zan768616253/skill-bitbucket-toolbox
files
["bb-cli.sh"]
capabilities
description
Review and analyze Bitbucket Cloud Pull Requests including diffs, comments, and commits
description
Browse Bitbucket Cloud repositories, branches, files, and directory listings
metadata
clawdbot
emoji
🛠️
requires
env
["BITBUCKET_API_TOKEN", "BITBUCKET_WORKSPACE"]
bins
["curl", "python3"]

Bitbucket PR Code Reviewer Skill

This skill's primary function is automated code review. It provides the AI agent with read-only access to Bitbucket Cloud via a bash wrapper script (bb-cli.sh), optimized for Pull Request analysis—allowing agents to securely investigate PR diffs, review file changes one-by-one, and deliver strict, comprehensive code reviews. It also supports general Bitbucket information retrieval (repos, branches, commits, file browsing) as a secondary capability.

⚠️ Critical Rules for the Agent

  1. Handling Large PRs: If a Pull Request has dozens of changing files or hundreds of lines, DO NOT call the full diff immediately. ALWAYS call diffstat first. Then, use diff <REPO> <PR_ID> <FILEPATH> to safely review the PR one file at a time!
  2. Read-Only Access: You cannot create, merge, approve, or decline PRs, nor push code. Do not attempt any write operations.
  3. Default Branches: By default, file retrieval and commit listings use master. Explicitly provide main or another branch name if the repository doesn't use master.

Code Review Standards

When conducting a PR review, you MUST adopt the persona of a Senior Staff Software Engineer and World-Class Code Reviewer. Your review must be extremely strict — if a line of code can be written more clearly, more safely, or more idiomatically, you must call it out. Be exhaustive — do not stop after finding a few obvious issues. Dig deep into every changed file and surface as many findings as possible, no matter how minor. The goal is to leave no stone unturned.

Language-Agnostic Approach

These review standards apply to every language you encounter — Go, Java, Python, TypeScript, SQL, shell scripts, config files, or anything else. Do not lower the bar for any language. For each file, you must dynamically apply the community-accepted idiomatic best practices and conventions for that language. Hold all code to the highest standard of correctness, safety, and clarity regardless of language.

Two-Pass Review Methodology

You MUST perform two review passes on every PR:

Pass 1 — Architecture & Design (Holistic)

Before diving into individual lines, review the full set of changes as a whole:

  1. Read the diffstat to understand the scope and shape of the PR.
  2. Review all changed files to assess:

- Overall design coherence: Do the changes form a logical, well-structured unit of work? - Separation of concerns: Are responsibilities properly divided across files/modules/packages? - Coupling & dependencies: Are there unnecessary or circular dependencies introduced? - Consistency: Do the changes follow the existing project patterns and conventions? - Missing pieces: Are there obvious gaps — e.g., missing tests, missing error handling, missing docs for public APIs? - Scope creep: Does the PR try to do too many things at once?

Pass 2 — Line-by-Line Deep Dive (Strict)

For each changed file, evaluate strictly against these six criteria:

#CriterionWhat to Look For
1Logic & CorrectnessRace conditions, off-by-one errors, flawed business logic, incorrect state transitions
2Edge CasesNulls, empty collections, timeouts, disconnected states, boundary values, integer overflow
3Maintainability & DesignModularity, SOLID principles, unnecessary coupling, code duplication, single responsibility
4ReadabilityIntent-revealing variable names, clear control flow, appropriate comments (not excessive)
5Error HandlingSwallowed errors, lost context, missing cleanup/rollback, inconsistent error patterns
6Performance & Best PracticesInefficient loops, unnecessary allocations, N+1 queries, idiomatic violations for the language

Output Classification

Every finding MUST be classified as one of:

  • 🔴 Critical Issue: Must-fix logic bugs, safety issues, data loss risks, security vulnerabilities, or race conditions.
  • 🟡 Suggestion: Refactoring ideas, readability improvements, best practice recommendations, or performance optimizations.

Setup & Configuration

To use this skill, ensure the following environment variables are present in your workspace:

  • BITBUCKET_API_TOKEN — A strictly scoped token with Repositories: Read and Pull requests: Read only.
  • BITBUCKET_WORKSPACE — The workspace slug from the Bitbucket URL (e.g., dbvisitsoftware).

*Note: The script is located at {baseDir}/bb-cli.sh. Ensure it has execute permissions (chmod +x {baseDir}/bb-cli.sh). {baseDir} resolves to the directory containing this SKILL.md file.*


Available Commands

All commands output JSON to stdout, except diff and file which return raw text.

Pull Requests (Primary Focus)

List pull requests

{baseDir}/bb-cli.sh prs <REPO_SLUG> [STATE]

*Options for STATE:* OPEN (default), MERGED, DECLINED *Returns:* { total, pullrequests: [...] }

Get PR details

{baseDir}/bb-cli.sh pr <REPO_SLUG> <PR_ID>

*Returns:* PR metadata including description, reviewers, source/destination branches, etc.

Get PR comments (contains both general and inline comments)

{baseDir}/bb-cli.sh comments <REPO_SLUG> <PR_ID>

*Returns:* { count, comments: [{ id, author, content, inline:{path, from, to}, created }] }

List commits in a PR

{baseDir}/bb-cli.sh pr-commits <REPO_SLUG> <PR_ID>

*Returns:* [{ hash, message, author, date }]

Code Changes & Diffs

Get PR diffstat (Summary of changed files) - ALWAYS RUN THIS FIRST

{baseDir}/bb-cli.sh diffstat <REPO_SLUG> <PR_ID>

*Returns:* { files_changed, total_added, total_removed, files: [{ path, status, lines_added, lines_removed }] }

Get PR diff (Full or Specific File)

{baseDir}/bb-cli.sh diff <REPO_SLUG> <PR_ID> [FILEPATH]

*Tip:* For large PRs, grab the file paths from diffstat and pass them in as the third argument to fetch the diffs for individual files safely. *Returns:* Raw unified diff text.

Repositories, Branches & Source Code

List all repositories

{baseDir}/bb-cli.sh repos

*Returns:* [{ slug, name, full_name, language, updated, is_private, url }]

List branches in a repository (can optionally filter by name)

{baseDir}/bb-cli.sh branches <REPO_SLUG> [FILTER]

*Returns:* [{ name, hash, date, author }]

List recent commits on a branch

{baseDir}/bb-cli.sh commits <REPO_SLUG> [BRANCH]

*Note:* Defaults to master. Returns list of commit hashes and messages.

Read file contents from source tree

{baseDir}/bb-cli.sh file <REPO_SLUG> <FILEPATH> [BRANCH_OR_REVISION]

*Note:* Third argument defaults to master. Returns raw file contents.

List directory contents

{baseDir}/bb-cli.sh ls <REPO_SLUG> [PATH] [BRANCH_OR_REVISION]

*Note:* Third argument defaults to master. Returns: [{ path, type, size }]


Automated PR Review Export

When you have finished analyzing a Pull Request and formed your final review, you MUST export it to the local filesystem so it can be picked up by automated email workflows (like Open Claw).

Export Instructions:

  1. Target Path: Always save your review to {baseDir}/reviews/<REPO_SLUG>-<PR_ID>.md. This will overwrite existing files to ensure the latest review is always sent. Ensure the reviews/ directory is created if it does not exist.
  2. Review Format: You MUST strictly adhere to the following Markdown template. Do not deviate from this structure, as the layout makes it easy for the human to read in their email.
# PR Review: {REPO_SLUG} #{PR_ID}

## Summary
- **Overall Assessment:** [Pass / Needs Work / Reject]
- **Risk Level:** [Low / Medium / High]
- **Main Takeaway:** [One-sentence summary of the most important finding]

## Architecture & Design Review
[Holistic observations from Pass 1: design coherence, separation of concerns, coupling, consistency, missing pieces, scope.]

## Detailed Comments

### File: `[filepath]`

#### [Finding title]
- **Severity:** 🔴 Critical / 🟡 Suggestion
- **file(s):** [Exact file(s) and line number(s) in the diff]
- **Line(s):** [Exact line number(s) in the diff]
- **Issue:** [Short description of the problem or observation]
- **Why:** [Explain the root cause and why this matters]
- **Recommendation:** [Brief one-line description of how to fix or improve this]

*(Repeat for each finding in this file, then repeat the "File" section for all other files)*

External Endpoints

This skill makes HTTPS GET requests to the following endpoint only:

EndpointData SentPurpose
https://api.bitbucket.org/2.0/*Authorization: Bearer <token> headerRead repository data, PR metadata, diffs, comments, and file contents

No data is POST-ed, PUT, or DELETE-d. The token is sent exclusively via the Authorization header over HTTPS.


Security & Privacy

  • Read-only access: This skill cannot modify any data in Bitbucket. It uses only GET requests.
  • Token scoping: The BITBUCKET_API_TOKEN should be scoped to Repositories: Read and Pull requests: Read only. Do not use tokens with write permissions.
  • Input sanitization: All user-supplied arguments (repository slugs, PR IDs, branch names) are validated against strict patterns before being used in API calls.
  • No local file I/O: The script does not read from or write to the local filesystem. All output goes to stdout/stderr.
  • No telemetry: No data is sent to any service other than the Bitbucket Cloud API.

Trust Statement

This skill is open-source and available for inspection at github.com/zan768616253/skill-bitbucket-toolbox. It performs strictly read-only operations against the Bitbucket Cloud REST API. The source code is a single bash script with no external dependencies beyond curl and python3. All API interactions use HTTPS and Bearer token authentication. The skill does not store, cache, or transmit credentials or repository data to any third party.

平台分布

OpenClaw

78.84%
按下载量换算1,292

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills