# Newsletter — drop-in blurbs (50 / 150 / 300 words)

Three lengths. Pick whichever fits the slot in your newsletter. Replace `getgodmode.dev/?ref=YOUR_SLUG` with your link. Disclosure line at the end is required for FTC compliance — keep it visible, don't bury it.

---

## 50-word blurb (sponsorship slot / "what I'm using" mention)

**Godmode for Claude Code (affiliate)** — adds an 8-layer execution loop on top of Claude Code: plan, sub-agents, parallel exec, auto-verify, self-fix, evolve. The verifier scores every output ≥0.92 or repairs it. Free Lite tier, paid from $99/yr. Try free → [getgodmode.dev/?ref=YOUR_SLUG](https://getgodmode.dev/?ref=YOUR_SLUG)

*Affiliate link. Doesn't change your price.*

---

## 150-word blurb (mid-issue "tool I'd recommend" feature)

**Tool I'd recommend this week: Godmode for Claude Code** *(affiliate)*

If your AI coding agent ships first attempts and you spend half your time re-prompting, Godmode is the missing layer. It wraps every Claude Code task in an 8-layer protocol — plan, parallel sub-agents, auto-verify, self-fix, skill evolution. Anything that doesn't score ≥0.92 goes back through the loop before it reaches you.

What changed for me in two weeks: 1-prompt completion rate climbed from ~25% to ~70%. I stopped operating Claude and started reviewing it.

Free Lite tier (same protocol, fewer skills). Paid plans start at $99/yr.

Try it free → [getgodmode.dev/?ref=YOUR_SLUG](https://getgodmode.dev/?ref=YOUR_SLUG)

*Disclosure: affiliate link, small commission to me at no extra cost to you.*

---

## 300-word blurb (deep-dive feature / "why I subscribed")

**Why I added Godmode to my Claude Code stack** *(affiliate)*

I've been running Claude Code daily for around six months. The model's good. The tool's good. But my "1-prompt completion rate" — tasks that finished cleanly without me re-prompting — sat stubbornly at about 25%.

The other 75% was me telling Claude what it broke.

Two weeks ago I tried Godmode, an execution skill stack that runs on top of Claude Code. It adds eight layers most agent flows skip: a plan phase, parallel sub-agents, automated output scoring, a self-fix loop, and an evolution layer that makes skills sharper the more you use them. Outputs that score below 0.92 get repaired before you see them.

The numbers shifted fast. By week two my 1-prompt completion was around 70%. The shift wasn't the model — it was the verification loop.

Three things stood out:

1. **The plan-first protocol.** Forces a task plan before any code is written. Fewer mid-task refactors.
2. **Sub-agents in parallel.** Most "agent" tools still run sequentially. Godmode fans out and reconciles.
3. **Skill evolution.** The skills you use most start producing tighter outputs. Week 4 me has different default outputs than week 1 me.

The Lite tier is free and runs the same protocol with fewer skills, so you can try the loop without paying. Paid plans start at $99/yr; Godmode+ at $149/yr adds the planner, sub-agents, and auto-verify layers explicitly.

Free Lite tier → [getgodmode.dev/?ref=YOUR_SLUG](https://getgodmode.dev/?ref=YOUR_SLUG)

*Affiliate disclosure: I get a small commission if you upgrade through the link above. Doesn't change your price. Lite tier is fully free, no card needed.*