/carbon
Test a project or verify a recent feature or bug fix with Jay: risk-based checks, reproducible findings and confidence in a visual report.
Jay · Your AI test managerStart & investigateWhen to use it
You have a change to validate but do not yet know where testing effort belongs.
Before you start
Give Jay the changed area and the user outcome that must still work. Check that the report distinguishes executed checks from source inspection and deferred risks.
Try this in your coding agent.
/carbon Review my changes in src/checkout and test the highest-risk purchase flow.Replace sample paths with your project. Slash names are shorthand; your host may use a plugin prefix or a skill picker. Invocation help →
ILLUSTRATIVE WALKTHROUGH
Checkout changes
An expanded teaching example, not a recorded or live execution. The report stays visible while the replay explains how to read it.
- 01Understand change
- 02Rank risk
- 03Run checks
- 04Reproduce
- 05Report
Coding-agent conversation
I’m reviewing checkout changes. I’ll keep the scope and permissions explicit.
I’m prioritizing incorrect totals and lost cart state before cosmetic differences.
Run checks: Review my changes in src/checkout and test the highest-risk purchase flow.
Here is the evidence trail to inspect: Cart: coffee + mug → apply MUG10 → remove mug → discount remains.
A discount survives after its eligible item is removed. Payment-provider behavior is outside this local example.
Example report
Illustration
TestBucks course screen with illustrative analysis. The screenshot supplies context; the demo does not establish a newly observed bug.
- Cart input
- Totals & state
- Checkout evidence
Illustrated outcome
A discount survives after its eligible item is removed.
Evidence to inspect
Cart: coffee + mug → apply MUG10 → remove mug → discount remains.
Still untested / not established
Payment-provider behavior is outside this local example.
Ready. Read the full example or play its steps.
What a useful result contains
A bounded claim
For checkout changes, state the target, revision, environment and actual work performed. Keep planned, executed, blocked and untested work separate.
An inspectable trail
Cart: coffee + mug → apply MUG10 → remove mug → discount remains. In a real run, retain the source, trace, screenshot or fixture behind the conclusion.
The qualification
Payment-provider behavior is outside this local example.
The next decision
Decide whether the evidence calls for a focused check, a reviewed fix or more context. Do not convert a plausible example into a verified finding.
Adapt the example to your project
- Replace checkout changes with a real feature, path or decision in your project.
- State the expected rule and its source. Include safe fixtures, constraints and a budget where appropriate.
- Ask for exact evidence and remaining gaps. Approve consequential actions separately.
/jay Review the /carbon result. Separate observed evidence, assumptions and untested scope, then recommend the next useful check.Where to go next
/carbon-more →/carbon-test →Content checked against the local command definition: skills/carbon/SKILL.md. Examples adapted from the homepage’s TestBucks walkthroughs. Availability is edition- and host-dependent.