<p align="center"> <img src="assets/images/banner.png" alt="Strategic Partner - Chief of Staff for Claude Code" width="100%"> </p>

[](CHANGELOG.md)

# strategic-partner

A strategic advisory skill for Claude Code (an installable add-on that extends Claude Code's behavior) that separates thinking from building. It thinks with you in one session — asking the right questions, challenging assumptions, framing problems before jumping to solutions. Then it packages implementation for fresh sessions where the full context window is available. Decisions persist. Context stays clean.

> **What's new** — **7.6.0** adds a Serena setup steward: it checks each user's real setup, offers a reversible fix when needed, binds Serena to the exact repository or worktree, and stops the dashboard from opening a browser tab automatically. See [CHANGELOG.md](CHANGELOG.md) for the full list and prior releases.

---

## The problem

AI coding assistants degrade as conversations grow. Every tool call, file read, and back-and-forth exchange pushes the original instructions further from the model's attention. By the time you're deep into implementation, the careful thinking from earlier has been diluted by hundreds of intermediate results.

Think of it like a meeting that started with a clear agenda but kept going for six hours. By hour four, decisions are being made on autopilot — not because anyone stopped caring, but because the original focus got buried under everything that came after.

Most workflows ignore this. You open one session, plan and build in the same window, and by the time you're in the weeds, decisions are being made mid-build with degraded instruction-following. When context fills up completely, everything is lost.

The strategic partner fixes this by enforcing a separation: persistent advisory context where decisions accumulate, and disposable execution context where clean context matters most.

---

## Who is th