# How Kent C. Dodds actually uses coding agents

Testing Library, Epic Web

**If you take one thing:** `/ship-pr`, Bugbot loop, merge, Discord. a cloud agent babysits its own PR through Bugbot and CI, merges low-risk changes, then posts a summary with the billed agent cost to his Discord.

- **Agent:** Grok Bot ([source, 2026-10-11](https://x.com/kentcdodds/status/2109131437006074355))
- **Agent:** Cursor (web) ([source, 2026-09-27](https://x.com/kentcdodds/status/2104285773574885610))
- **Model:** Grok 4.5 ([source, 2026-08-27](https://github.com/kentcdodds/kcd-skills/blob/ab6303e3b22fbbeec665f9a62ee654f379c8c600/skills/orchestrate/SKILL.md#L25-L28))
- **Model:** Fable ([source, 2026-08-27](https://github.com/kentcdodds/kcd-skills/blob/ab6303e3b22fbbeec665f9a62ee654f379c8c600/skills/orchestrate/SKILL.md#L25-L28))
- **Review:** Trusts gates and AI reviewers, reads the outcome, not the diff ([source, 2026-09-24](https://kentcdodds.com/blog/you-own-the-outcome))
- **At once:** a dozen ([source, 2026-08-27](https://github.com/kentcdodds/kcd-skills/blob/ab6303e3b22fbbeec665f9a62ee654f379c8c600/skills/orchestrate/SKILL.md#L41-L42))
- **Harness:** Cursor Cloud Agents open PRs; CI gates and Bugbot review; ship-pr skill merges and reports to Discord ([source, 2026-10-06](https://github.com/kentcdodds/kentcdodds.com/blob/76fdb8987b95973066daf59d2807c2bb9bed722d/docs/agents/project-context.md#L184-L189))
- **Rules file:** AGENTS.md as a tiny map to docs/agents/ ([source, 2026-08-17](https://github.com/kentcdodds/kentcdodds.com/blob/b61c7ff4200c4aa0e35326d1a7baeb8a51d86673/AGENTS.md#L3-L4))
- **Skills:** ask-kent, review-and-recommend, orchestrate, visual-recap, ship-pr, prefer-local-kody-execute ([source, 2026-10-01](https://github.com/kentcdodds/kcd-skills/tree/ab6303e3b22fbbeec665f9a62ee654f379c8c600/skills))
- **Hooks:** pre-push test gate (npm run test:push), composed with Cursor Cloud agent hooks ([source, 2026-10-06](https://github.com/kentcdodds/kody/blob/65ac15100a5c1dde9b6a8387404f0bd775602a01/docs/contributing/cloud-agents.md#L75-L81))
- **MCP servers:** kody ([source, 2026-10-01](https://github.com/kentcdodds/kcd-skills/blob/ab6303e3b22fbbeec665f9a62ee654f379c8c600/skills/prefer-local-kody-execute/SKILL.md#L11-L12))

## In their words

> Grok bot (my main driver for everything)

> In the last couple weeks I've had Cursor make me two native Android apps for personal software I wanted because it's that easy

> Prefer **Grok 4.5** (`grok-4.5-fast-xhigh`) for coding sub-agents; `composer-2.5-fast` for mechanical work.

> **Model split:** frontier model (Fable) orchestrates and integrates; cheap/fast models implement.

> **Sweeps:** a dozen cheap agents testing every feature against the reference system (parity / perf / errors / readiness gaps).


Archetype (our label): Orchestrator. Newest source: 2026-10-11.
Card: https://www.prehistoriq.com/stacks/kentcdodds/ · Data: https://github.com/prehistoriq/stacks/blob/main/data/stacks/kentcdodds.json