How to use Claude to add expression output to a dev tool
After abandoning an earlier attempt, a developer used Claude to add expression-and-value output to gg within hours. They also reported a virtual development console optimized for coding agents.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Help extend [debugging tool repository] so its output includes both [source expression or variable name] and [evaluated value], similar to [reference tool]. Explore implementation alternatives with me and propose a plan before making the change.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- Leftium
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reported adding expression-and-value output within a few hours and having a virtual development console optimized for coding agents.
- Human involvement
- The author describes specification work and exploring alternatives, but does not report specific code-review or testing steps for gg.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Start with the hand-written gg project and attempt expression-and-value output.
- Work with Claude to add the previously abandoned feature within a few hours.
- Extend gg with a virtual development console optimized for coding-agent interaction.
Leftium described using Claude to implement ideas they had previously abandoned or would not have attempted alone. One example was gg, a project that began as entirely hand-written code. The developer wanted it to print the expression from the source code alongside its value, similar to Python's icecream tool.
They had already tried to add that behavior and given up. With Claude's assistance, they reported adding it within a few hours. The goal was to expose both the value and the variable name or expression, rather than the value alone.
The author also reported that gg now had a virtual development console optimized for interaction with coding agents. They described this as a substantial extension of its existing open-in-editor feature and something they probably would not have attempted without Claude.
Across their work, the author said specification effort varied with feature complexity and included research and exploration of alternatives. Sometimes the only specification was Claude's plan in the chat. The post does not describe gg's implementation or code-review process.
Who it's for: developers extending debugging tools who have a concrete behavior they previously struggled to implement.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · Leftium
Frequently asked questions
What output behavior did the developer add to gg?
The author wanted gg to print the source expression alongside its value, including the variable name or expression rather than only the resulting value.
How long did the reported feature take?
The author says Claude helped add the expression-output feature within a few hours after an earlier attempt had been abandoned.
Does the post explain how the feature was implemented?
No. It reports the behavior and approximate time but does not describe the underlying implementation or review process.
More details
About this account
Claude assisted with feature development and planning; the author describes varying specification detail depending on complexity.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=47033978. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: Leftium on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude · gg
Developer tools · Debugging · Feature development
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.