project-context

project-context

A local, provider-neutral MCP server for repository-scoped issue handling. It provides a guarded interface to Linear, GitHub Issues, GitHub Projects v2, and Jira Cloud, with preview/apply safety and host-local configuration.

Category
访问服务器

README

project-context

project-context is a local, provider-neutral MCP server for repository-scoped issue handling. It gives agents one guarded interface to Linear, GitHub Issues, GitHub Projects v2, and Jira Cloud while keeping project mappings and credentials outside source repositories.

The server resolves the current Git repository, selects its configured issue provider, verifies the authenticated account, and separates every write into a preview and a single-attempt apply step. Pull requests, releases, repository administration, Git authentication, and commit signing are intentionally out of scope.

Quick start

Requirements: Node.js 22 or newer on Linux or macOS.

VERSION=0.1.0
npx -y --package="@sandovalrr/project-context-mcp@$VERSION" project-context setup
npx -y --package="@sandovalrr/project-context-mcp@$VERSION" project-context doctor

The first command creates secure, host-local starter files without overwriting anything that already exists:

~/.agents/config/project-context/projects.yaml
~/.agents/config/project-context/credentials.yaml
~/.agents/config/project-context/templates/
~/.agents/config/project-context/secrets/
~/.local/state/project-context/

Copy and adapt the sanitized examples in examples/, validate the result, and then register the project-context-mcp stdio command in your MCP client. The complete Codex, Claude, Zed, and VS Code configurations are in docs/installation.md.

Safety model

  • Host-local YAML determines the repository, provider, target, and expected identity. Missing or ambiguous context fails closed for writes.
  • Credentials resolve from files, environment variables, OS keychains, or argv-only commands. Literal secrets in YAML are rejected.
  • Provider calls are restricted to fixed HTTPS origins, reject redirects, cap response sizes and timeouts, and never automatically retry writes.
  • Linear issue operations use a fixed allowlist of Linear's hosted MCP tools; project-context never proxies the upstream server's broader tool catalog.
  • Prepared writes expire after ten minutes, are encrypted with AES-256-GCM, and can be claimed only once. An uncertain apply result becomes indeterminate and is never replayed automatically.
  • Mutation audit records contain metadata only, are mode 0600, and rotate locally. The project has no telemetry.

Read docs/threat-model.md for trust boundaries and residual risks, and docs/routing-and-safety.md for operational behavior.

Commands and MCP tools

project-context --help
project-context setup --guided
project-context config validate
project-context resolve --cwd /path/to/repository
project-context issue list --status open --status in_progress
project-context issue list --parent linear:ENG-42
project-context issue user search "John Smith"
project-context issue capabilities
project-context issue option search labels security
project-context issue option search cycle current
project-context issue get linear:ENG-42 --include-relations
project-context issue comment list github:#42 --limit 20
project-context issue prepare create --title "Child task" --parent ENG-42
project-context doctor
project-context audit list
project-context skill status
project-context integration manifest --client codex

The stdio server exposes eleven tools:

  • resolve_project_context
  • list_issues
  • search_issues
  • list_users
  • search_users
  • search_issue_options
  • get_issue_capabilities
  • get_issue
  • list_issue_comments
  • prepare_issue_change
  • apply_issue_change

Every tool has an input and output schema. Starting the server has no setup side effects; when host configuration is absent, tools return an actionable structured error.

Linear additionally supports target-scoped subissues, due dates, estimates, cycles, milestones, blocking/related/duplicate relationships, comment replies and edits, archived listing, and relation-expanded reads. SLA fields, permanent deletion, and non-issue Linear features are not exposed.

GitHub additionally supports repository issue types and milestones, native parent/subissue hierarchies, blocking dependencies, duplicate relationships, comment edits, direct-subissue listing, and relation-expanded reads. GitHub custom-field mappings, generic related-to relationships, threaded comment replies, pull requests, and repository administration are not exposed.

integration manifest prints the provider-neutral command definition by default. Pass --client codex, claude, zed, or vscode to emit native, ready-to-paste configuration pinned to the installed package version. Add --json when automation needs an envelope containing the client, format, and configuration.

The optional project-issues agent skill is packaged but never installed as an npm lifecycle side effect:

project-context skill status
project-context skill install

Use --replace only after reviewing status; replacement creates a timestamped backup.

Documentation

License and attribution

This project is available under the MIT License, copyright 2026 Richard Sandoval. The copyright and permission notice must remain with copies or substantial portions of the software.

If this repository is used as the starting point for another project, a visible acknowledgment is appreciated:

Based on project-context, created by Richard Sandoval.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选