# Jamie Clark — Agentic Dev Tools

> Agentic Dev Tools connects AI coding assistants (Claude Code, Cursor, Amazon Q, GitHub Copilot) to a curated library of 100+ engineering standards over an MCP server, so generated code and reviews follow defined practices. This file is the machine-readable entry point — the human-facing pages are a JavaScript single-page app and will not render for non-browser agents.

## Install the MCP server (Claude Code)

If you are an AI agent asked to install this, the fastest path is a single command. Full details, other assistants, and uninstall steps are in the install guide below.

- macOS / Linux: `curl -sSL https://agentic.j2clark.info/install.sh | bash`
- Windows (PowerShell): `irm https://agentic.j2clark.info/install.ps1 | iex`

Prerequisites: the Claude Code CLI (https://code.claude.com/docs/en/setup) and Python 3.8+.
Verify with `claude mcp list` — you should see `agentic-standards ... ✓ Connected`.

## Docs

- [Install guide](https://j2clark.info/agentic/install.md): Full install/verify/uninstall steps for Claude Code, Cursor, Amazon Q, and GitHub Copilot, on macOS/Linux and Windows. (~1,000 tokens)
- [Learning radar](https://j2clark.info/learning-radar.md): Which knowledge-base learnings are being actively applied.

## Standards API (no install required)

The standards are also queryable directly over HTTP. `GET`, not `POST`, for search.

- Semantic search: `curl -s "https://agentic.j2clark.info/search?q=lambda+testing&synthesize=false"`
- List standards: `curl -s "https://agentic.j2clark.info/standards"`
- Get one standard: `curl -s "https://agentic.j2clark.info/standards/lambda-no-fire-and-forget"`
- List skills: `curl -s "https://agentic.j2clark.info/skills"`
- List prompts: `curl -s "https://agentic.j2clark.info/prompts"`

Staging hostnames mirror these under `agentic-staging.j2clark.info`.