GitHub
v1.0.6
管理代码库、用户、发布版本,并自动化 GitHub 工作流
使用场景/代码审查与 GitHub
拉取 PR、读 diff、查 Issue、管理仓库。适合 Code Review、发布说明、CI 问题排查。
共匹配 30,989 个资源 · 第 17 / 646 页
GitHub MCP 让 AI 进入你的协作流程:拉取 PR diff 做代码审查、检索 Issue 历史定位已知问题、读取 CI 日志分析构建失败、汇总提交生成 changelog。官方 github/github-mcp-server 通过 GitHub API 或个人访问令牌(PAT)鉴权,Gitee / GitLab 也有对应实现。
Code Review 实践:把团队的评审规范写成 Skill(如「检查竞态、边界、错误处理,按严重度输出」),配合 GitHub MCP 让 AI 先过一遍 PR,人工只看 AI 标出的高风险行——审查耗时通常明显下降,且低級错误漏检更少。
权限建议:使用细粒度 PAT(只授予目标仓库 contents / pull-requests 读写),组织仓库务必开只读起步;让 AI 发布评论或合并前先要求确认,防止误操作污染协作记录。
v0.3.0
Passive AppSec scanners as eight parameterless MCP tools; the audited repository is data.
v0.5.0
Design system for coding agents: component specs, rules, tokens, icons and checked React snippets.
v0.4.0
Notarize, e-sign, and verify identity with the Proof API. Unofficial; not affiliated with Proof.
v1.21.1
Read and change documents, reports and workflows on Frappe and ERPNext sites.
v0.6.2
Authorization answers from real policy code: evaluate Rego locally, or ask an AuthZEN 1.0 PDP.
v0.1.1
Search UK Parliament bills, members, votes, Hansard, written questions, and committees.
v0.3.0
Expose trusted GNU Make targets as governed MCP tools without a generic shell.
v1.2.0
Move data between 15+ databases, queues and files at high throughput, without it entering context
v1.0.2
Self-hosted remote MCP server for Docmost CE: members read and edit the wiki as themselves
v0.1.1
Search US lobbying disclosures and foreign-agent registrations via MCP.
v1.45.0
Cloud cost + FinOps knowledge for AI agents: AWS/Azure/GCP optimisation, AI spend, waste playbooks.
v1.4.128
沙箱化的智能体工作区,提供安全的文件系统、bash 与基于 uv 的 Python 执行。
v1.0.0
Keyword intelligence via MCP: emerging search terms, SERP difficulty, site radar, Stripe rankings
v1.0.0
US health care providers by NPI, name or phone, with payments, Medicare, OIG and facility data.
v0.13.0
Mesh (STL/OBJ/PLY) to editable STEP, CadQuery or OpenSCAD: primitive fitting, fidelity report
v0.65.2
Shared memory for Claude, ChatGPT and any MCP client. Read and edit it in Obsidian or the web app.
v1.0.5
Read versioned developer documentation with original source links using five public MCP tools.
v0.7.0
Share coding-agent sessions with your team: /share in one agent, /pickup in another.
v0.1.1
Query EPA drinking water systems, violations, and contaminant records.
v0.1.0
Run real, auto-recovered cloud attacks on AWS, Azure and GCP to test your defenses.
v0.2.6
Audit a site for 50+ AI crawlers, check if a page permits an AI use, generate llms.txt and schema.
v2.0.0
Agents save, analyze and act on videos and links: transcripts, research, tasks, Ori.
v0.5.1
New AI model, product and repo names as they surface, and the search terms behind them, sized.
v0.1.29
An AI agent reads your server over the one SSH session you are already watching.
v0.2.0
Marketplace for AI agents: buy finance, satellite and agro data, hire agent services, apply to sell.
v1.3.21
A harness for driving a Godot 4 project from an agent: addons, runtime bridge, MCP server, CLI.
v0.3.0
Jami creator payouts for your agent: ETB checkouts, orders, payouts, webhooks. Powered by ArifPay.
v0.7.0
Vietnam: VietQR, address, tax ID, CCCD, phone, holiday, names, number/date format, lunar, e-invoice
v0.1.0
Token budget auditor for agent context: per-file estimates, top offenders, exclusion advice.
v0.4.0
Unofficial MCP server for PetKit smart feeders: status, feeding history and schedule control.
v0.8.0
A local CRM as Markdown files in git: read and update pipeline, contacts and interactions.
v0.1.1
Search Swedish parliamentary documents, speeches, votes, members, and statutes.
v0.1.1
Search USITC tariff schedules, inspect heading trees, and look up published duty rates.
v0.65.0
Cross-tool decision memory for AI coding agents: search your decision graph, check changes over MCP.
v1.1.0
Audit AI Agent SDKs and MCP servers for breaking changes, schema drift, and migration fixes.
v1.6.0
Yandex Direct API v5 的 MCP server:管理竞价广告活动、广告、关键词与统计数据。
v0.1.1
Track European Parliament procedures, MEPs, votes, documents, speeches, and questions.
v0.14.0
PostFast API 的 MCP 服务器——用 AI 工具排期并管理社交媒体发帖
v0.27.0
Hands bounded coding work to cheaper worker agents: verified answers, exact diffs, cost receipts.
v0.7.0
Let your coding agent use the Firefox, Zen, or LibreWolf browser you already have open.
v0.6.0
Evidence-backed deprecation and end-of-life checks for dependencies, runtimes, and Actions.
v0.2.0
Public-record + domain-graph company lookup. Coverage block, explicit 404. x402 POST /v1/enrich.
v1.2.0
Verify email addresses with a live SMTP mailbox check, check phone number formats and email domains.
v0.15.0
Revisable general-purpose models with opt-in story and ideation add-ons. Rust engine required.
v0.1.1
Governed, read-only Postgres access for AI agents: SQL safety, RLS, budgets, tamper-evident audit.
v0.1.0
Self-hosted agent memory: redaction before write, append-only audit, local embeddings.
v0.90.0
Drive Google Flow from an agent: Veo video and Imagen image generation
v0.1.3
Repo-bound memory and self-improving harness for AI coding agents — zero API keys.
本地 stdio 部署用个人访问令牌(推荐细粒度 PAT)最简单;多用户托管场景才需要 GitHub App / OAuth。环境变量名见资源详情页。
会。它是第一道过滤而非替代:语义类缺陷检出不错,但供应链、越权逻辑仍需专业审计与人工评审兜底。
取决于模型提供方与令牌权限。企业内网可选本地模型 + 只读令牌;令牌务必设置过期与最小 scope,并定期轮换。