菜单
Put pctx between agents and tools so MCP servers and custom tools can run through a code-oriented execution layer instead of one-off tool calls.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use Vocode to compose transcription, LLM, speech synthesis, and telephony components into reviewable real-time voice-agent workflows.
Use Claude Legal Skill to guide an agent through contract risk review, CUAD-style clause checks, market benchmarks, and lawyer-ready redline suggestions.
Use SWE-ReX when a coding agent needs a consistent runtime interface for local, containerized, or remote shell sessions with command output, exit codes, interactive tools, and parallel execution.
Use Text Extract API when an agent needs to turn PDFs, Office files, or images into Markdown or structured JSON with local OCR, optional Ollama models, PII removal, and queued batch processing.
Route model calls through Helicone, inspect costs, latency, traces, prompts, and evaluations, then review changes before they ship.
Launch an OpenAI-compatible OpenLLM server for a chosen open model, point an agent runtime at it, and compare behavior before production use.
| name | Run agent tool calls through pctx Code Mode and unified MCP |
| slug | run-agent-tool-calls-through-pctx-code-mode-and-unified-mcp |
| description | Put pctx between agents and tools so MCP servers and custom tools can run through a code-oriented execution layer instead of one-off tool calls. |
| github_stars | 264 |
| verification | security_reviewed |
| source | https://github.com/portofcontext/pctx |
| author | Port of Context |
| publisher_type | organization |
| category | Developer Tools |
| framework | MCP |
| tool_ecosystem | {"github_repo":"portofcontext/pctx","github_stars":264,"npm_package":"@portofcontext/pctx","npm_weekly_downloads":285} |
Put pctx between agents and tools so MCP servers and custom tools can run through a code-oriented execution layer instead of one-off tool calls.
pctx CLI, Python SDK or npm package, MCP-compatible client or OpenAI Agents SDK
Install the pctx CLI and the client library that matches your agent stack:
brew install portofcontext/tap/pctx
npm i -g @portofcontext/pctx
pip install pctx-client
Use the Python client when building Python agents that call pctx as an HTTP service. Use the npm package or Homebrew CLI when you want pctx available as a local command for Code Mode or MCP setup. Existing Homebrew installs can be updated separately with brew upgrade pctx; do not treat that as the first-install path.
Basic setup:
pctx start
pctx mcp init
Run pctx as the boundary between your agent and registered tools/MCP servers. Keep the MCP server list explicit, review which tools are exposed through Code Mode, and avoid giving the bridge broader filesystem or network access than the workflow needs.