# Bootstrapping Coding Agents: The Specification Is the Program

arXiv: https://arxiv.org/abs/2603.17399 · HTML: https://mlsanity.com/paper/2603.17399/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 4 of its 13 references that are in the map.
- First posted on arXiv 18 Mar 2026.

## Builds on (most cited first, 4 of 4)

- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [Spec-Driven Development:From Code to Contract in the Age of AI Coding Assistants](https://mlsanity.com/paper/2602.00180/) (2026)
- [Galapagos: Automated N-Version Programming with LLMs](https://mlsanity.com/paper/2408.09536/) (2024)
- [Debootstrapping without Archeology: Stacked Implementations in Camlboot](https://mlsanity.com/paper/2202.09231/) (2022)

## Built on by (most cited first, 1 of 1)

- [From Prompt to Harness: Coderlet from Scratch](https://mlsanity.com/paper/2608.09480/) (2026)

## Abstract (arXiv)

A coding agent can bootstrap itself. Starting from a 926-word specification and a first implementation produced by an existing agent (Claude Code), a newly generated agent re-implements the same specification correctly from scratch. This reproduces, in the domain of AI coding agents, the classical bootstrap sequence known from compiler construction, and instantiates the meta-circular property known from Lisp. The result carries a practical implication: the specification, not the implementation, is the stable artifact of record. Improving an agent means improving its specification; the implementation is, in principle, regenerable at any time.