llm-coding-agent 0.1a0

Simon Willison · Simon Willison · 2026-07-02

Simon Willison releases llm-coding-agent 0.1a0, an open-source Python coding agent library generated by Claude Fable 5 via red/green TDD, implementing six file and command tools accessible via CLI and an unsolicited Python class API.

Open original ↗

Appears in

Claude Fable 5: Model Update, Safety Profile, Benchmarks, and Subscriber Trial Rollout

Extraction

Topics: coding-agentsllm-toolingopen-source-aiai-generated-code

Claims

Claude Fable 5 via Claude Code for web generated a complete coding agent Python library from a two-prompt specification using test-driven development.
The resulting library implements six tools: edit_file, execute_command, list_files, read_file, search_files, and write_file.
The AI created an unsolicited CodingAgent Python class API that Willison found useful despite not requesting it.
The library is available on PyPI as an alpha release and can be run without installation via uvx.
GPT-5.5 reasoned correctly that SwiftUI is unsuitable for a CLI and built an ASCII art time display app as a corrective alternative.

Key quotes

I've shipped a slop-alpha to PyPI, so you can run the new agent like this: uvx --prerelease=allow --with llm-coding-agent llm code

It also presents a Python API based around a CodingAgent(model='gpt-5.5', root='/path', approve=True).run('Fix the failing test in tests/test_parser.py') class which I didn't ask for but I'm delighted to see implemented.