# Lauren Tan’s pstack: verify first, then parallelize

> A quick take · Lauren Tan / pstack

---

LLMS index: [llms.txt](/llms.txt)

---

> The most critical skill to have in your toolbox is a high quality verification skill.
>
> Go faster with parallelism with Cloud Agents instead of worktrees
>
> — Lauren Tan · Source: [The Complete Guide to pstack Pt. 1](https://x.com/poteto/status/2094457600259842065); further reading: [Pt. 2](https://x.com/poteto/status/2097732320606507506), [Loops You Can Trust](https://x.com/poteto/status/2069824386283319343); [pstack Cursor plugin](https://github.com/cursor/plugins/tree/main/pstack)

She emphasizes the importance of verification, and why she uses Cloud Agents to parallelize development rather than git worktrees. (X really isn't a good place to publish or read articles.) In principle, these agentic development skills should be built into the harness; there shouldn't be a need to install them separately.

> The art of supervising someone smarter than you
>
> One technique I like to use is the indirect prompt. Instead of telling the agent exactly what I want, I try to draw it out of the agent instead - in its own words.
>
> This accomplishes three things:
>
> First, it forces the agent to compress a noisy conversation into a structured problem statement. Second, it lets me catch misunderstandings immediately. If the agent fixates on a red herring in the thread, I can correct it quickly before it starts writing any code.
>
> And third, I haven't potentially led it down the wrong path by stating my own assumptions and hypotheses which could be incorrect or limit what the agent could otherwise achieve.
>
> — Lauren Tan · [The Complete Guide to pstack Pt. 2](https://x.com/poteto/status/2097732320606507506)

This is great advice: don't rush to give instructions; first assess the agent's understanding of the problem.