Unit test suite for an existing codebase
Tests generated for your untested modules, each checked to fail for the right reason.
Diego FerreiraFull-stack developer, automation and dataAbout this gig
An agent writes the tests; I check that each one fails when the code is broken, delete the ones that test nothing, and wire the suite into CI.
You get coverage numbers before and after.
What is included
- Tests for up to 20 modules
- CI setup
- Coverage report
Add-ons
- End-to-end tests for three flows+$300 · +3 days
AI disclosure
How this work is made.
AI Involvement
AI-led
83%
A neutral measure: higher means more AI shaped the work, not worse work. Calculated from the stages below, weighted by how much each one counts. The seller cannot type it in.
Stage by stage
| Stage | Role of AI | Weight |
|---|---|---|
| Concept / ideation | AI produced it; the seller directed and curated | 15% |
| Research | AI output delivered essentially as-is | 10% |
| Production / drafting | AI output delivered essentially as-is | 40% |
| Refinement / editing | AI produced it; the seller directed and curated | 25% |
| QA / final review | AI assisted; the seller produced the output | 10% |
Tools, and what they are used for
- Claude CodeOpus 5
Writing the tests
What the seller does
Checking every test can fail, pruning weak ones, and the CI setup.
Declarations
- Source material
- Mixed
- A person reviews every deliverable
- Yes
- Prompt logs or source files on request
- Yes
- Your material and AI tools
- Shared only with tools that don't keep it
- Disclosing AI use to your own client
- Not needed
- Tool licences allow commercial delivery
- Yes
Staff verification
Not verified yet. Verification is a closer check than the review every gig passes before it goes live.
AI Involvement over time
Every change to the figure is kept, so a gig cannot quietly change how it is made.
- AI-led83%
What the seller needs from you
- What should it do, and what does done look like?Text · Required
- Link to the repository or the tools involvedLink · Optional
About the seller
Diego Ferreira
Full-stack developer, automation and data
I build web apps, glue SaaS tools together and clean up data, mostly for small teams without an engineer. I write code with agents every day and review every line they produce.
- AI tools they work with
- Claude CodeAgentic coding on whole features, with my review
- CursorDay-to-day editing and refactoring
- GitHub CopilotCompletions in the editor
- n8nWorkflow automation with AI steps
- ZapierIntegrations for non-technical teams
- How they work with AI
Coding agents write a lot of my first drafts. I design the architecture, review every diff, and test against real data before anything ships. Audits and reviews are done by hand, because judgement is the product there.
- Skills
- TypeScript
- Next.js
- Python
- SQL
- n8n
- Zapier
- Languages
- Portuguese · Native / bilingual
- English · Fluent
- Spanish · Conversational