How to use Cursor to prototype a dependency checker
The team released shouldiuse.dev after an internal AI-assisted experiment, providing public GitHub dependency checks, stored results, generated summaries and authenticated MCP access.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Prototype a dependency-health website for [repository scope] in Cursor. Use [frontend framework] and plan a Go backend around OpenSSF Scorecard, GitHub community metrics and stored results. Draft a concise JSON recommendation prompt that incorporates the check evidence and [FAQ questions].
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- louis_w_gk
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The team released shouldiuse.dev after an internal AI-assisted experiment, providing public GitHub dependency checks, stored results, generated summaries and authenticated MCP access.
- Human involvement
- Three people built prototypes and worked on merging their ideas. The source does not describe a formal review process for generated code.
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
- Created separate frontend prototypes using v0, Lovable and Cursor.
- Worked to merge prototype ideas despite differing frameworks and versions.
- Built a Go backend with Scorecard and experimented with Copilot.
The shouldiuse.dev team describes its first professional experiment with heavily LLM-assisted coding. The goal was to make dependency-health information accessible through a website built around OpenSSF Scorecard, which the team already used.
Three people independently created frontend prototypes using v0, Lovable and Cursor. Although the prototypes appeared quickly, combining their ideas proved difficult because they used different frameworks and versions. The team reports that frontend detail work and small adaptations consumed the most effort.
In parallel, they wrote a Go backend using the Scorecard library and deliberately experimented with Copilot, different models and prompts. They added community metrics from the GitHub API and generated textual summaries with OpenAI. The recommendation prompt combined role and format instructions, Scorecard results, community data and the questions shown in the website’s FAQ. Results were stored in PostgreSQL for subsequent retrieval.
The team also added an authenticated remote MCP server so coding assistants could access dependency checks from an IDE. At publication, the service supported public GitHub repositories and required a personal access token when requesting checks.
Who it's for: Teams exploring AI-assisted prototypes who can reconcile divergent implementations and evaluate dependency evidence themselves.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · louis_w_gk
Frequently asked questions
Was Cursor used for the entire project?
No. One frontend prototype used Cursor, while others used v0 and Lovable. The Go backend work involved Copilot, and OpenAI generated summaries.
What information went into the recommendation prompt?
The team supplied role, capability, limitation and JSON-format instructions, Scorecard results, additional community data and the questions displayed in the FAQ.
More details
Remaining steps
- Added GitHub metrics, OpenAI summaries, result storage and an authenticated MCP server.
About this account
The agent generated code from the author’s instructions, with the author directing the work.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=44399666. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: louis_w_gk on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Cursor · v0 · Lovable · Copilot · OpenSSF Scorecard · OpenAI · Go · PostgreSQL
AI coding
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.