How to use Claude to fix a JMAP mail server
After Claude generated a Rust JMAP server, roughly 100 conformance tests failed. A subsequent repair request led to a fully passing suite after about 45 minutes without further prompting.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Run [conformance suite] against [server repository]. Investigate each failure, fix the underlying implementation bugs and rerun the suite until it passes. Preserve the protocol requirements and report any behavior the suite does not cover.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- josephg
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The generated Rust JMAP server initially failed approximately 100 tests; after Claude's repair session, the full conformance suite passed.
- Human involvement
- The author ran the conformance suite, identified the failures and reports that the complete suite subsequently passed.
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
- Create a JMAP conformance suite with Claude Code and experts.
- Ask Claude to implement a simple JMAP-only server in Rust.
- Run the conformance suite against the generated server.
The author used Claude Code and experts to create a conformance suite containing more than 300 tests for JMAP, an email protocol. They then tried a separate implementation exercise: asking Claude to build a simple JMAP-only mail server in Rust.
After receiving the server implementation, the author ran the conformance suite against it. Approximately 100 tests failed. Rather than supplying a succession of individual debugging requests, the author asked Claude to fix all the bugs identified by the suite. Claude worked for about 45 minutes without further prompting during that interval.
The reported result was a fully passing conformance suite. This gives the account a concrete endpoint: the generated server initially failed many checks, and a subsequent autonomous repair session brought it into agreement with the supplied tests. The post does not establish production readiness, deployment or behavior outside the suite's coverage. Its demonstrated result is specifically the transition from those initial failures to passing the complete test suite.
Who it's for: Developers with executable protocol requirements that can guide an agent's debugging work.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · josephg
Frequently asked questions
What gave Claude a concrete debugging target?
The author ran a JMAP conformance suite containing more than 300 tests. Approximately 100 failed against the generated Rust server.
Did the author keep prompting during the fixes?
The author says no further prompting was needed during the roughly 45-minute repair session, after which the complete conformance suite passed.
More details
Remaining steps
- Ask Claude to fix every bug exposed by the failing tests.
About this account
After receiving the repair request, Claude worked for about 45 minutes without additional prompting.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=47332444. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: josephg on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Code · JMAP conformance test suite · Rust
Rust · JMAP · Conformance testing · Autonomous debugging
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.