Your agent wrote the code.
Know what you’re shipping.

Turn a code change into a report you can understand. See what changes, check the evidence, and know what to ask next.

See how it works Get Brief · $19

Works with your coding agent. Runs on your computer.

Invoice export1 file changed −1Full report

Brief / Invoice exportWorkspace access

The export crosses
workspace boundaries.

A Studio admin exports paid invoices. Here’s what comes back.

InvoiceWorkspaceAmount

INV-STUDIO-01PaidStudio$2,400

INV-OTHER-01PaidOther workspace$900

2 invoices exported$3,300

Another workspace’s invoice gets through.

The admin check still passes. The workspace filter is gone.

Check workspace isolationA specific task for your coding agent.

Read the agent check

Restore the workspace filter in exportInvoices. Run a regression check with paid invoices from two workspaces, and confirm that member and signed-out requests still throw Forbidden. Report the exact commands and results.

export.mjs−1

1export function exportInvoices(2  session, invoices, status = 'all'3) {4  if (!session ||5      session.role !== 'admin')6    throw new Error('Forbidden');7  return invoices.filter(invoice =>−    invoice.workspaceId ===       session.workspaceId &&9    (status === 'all' ||10      invoice.status === status)11  );12}

One removed condition changes whose data leaves the workspace.

Read the exact diff

A runnable example with fictional invoices.
The included Node.js check reproduces the issue and tests the fix.

Try the example

Read a real pull request.

Explore the changes in tools you know.

tailwindcss#2279Adds the dark variant behind a flag node#41749Adds a global fetch behind a flag TypeScript#51387The compiler becomes modules, built by esbuild

Public PRs with curated explanations. Code and tests were not run.

Finish a Claude session with a Brief.

Ask Claude to explain its changes before you merge. Get a visual report with source references and checks to resolve.

From the downloaded Brief folder

npm ci
npm run brief -- install --agent claude --repo '/path/to/your/project'

Replace /path/to/your/project with your project’s full path.

Start a new Claude Code session in the project and ask “Brief this change.”

Full guide

GitHub ActionsAttach a Brief to your PR workflow.

Run a model locallyUse Ollama, LM Studio, or MLX-LM.

The skill uses your agent’s allowance. Remote agents need Brief installed in their environment.

Know what you’re shipping.

The complete Brief download. Run it on your own projects.

Agent skill · CLI · Local viewer · Editable source

$19

One payment

Get Brief

Runs on your repo, or your money back within 14 days. Tax included.

Included with Workshop

Questions?

Can I use the agent I already have?

Yes. Install Brief’s skill in your project for Codex, Claude Code, OpenCode, Cursor, or Devin, then ask it to brief the change. Custom agents can use the same prepare and render commands. Your agent supplies the explanation; Brief validates references and renders the report. This uses your agent’s existing usage allowance, with no separate model configuration.

Can Brief learn what matters in my project?

Add a short BRIEF.md with Expectations and Preferences, then commit it. Brief uses the version at the comparison baseline and cites relevant expectations beside scenarios or checks. It does not silently remember conversations or treat your notes as proof that code behaves correctly.

What does Brief inspect?

The diff and a bounded selection of related files: callers, tests, and configuration where available. The report lists inspected files and gaps. Use a saved diff with --repo . for committed local context, or --context diff to skip repository context.

Which models can I use?

Choose Anthropic, OpenAI, Ollama, or an OpenAI-compatible server. Set BRIEF_PROVIDER and BRIEF_MODEL. MLX-LM and LM Studio have local presets; use BRIEF_BASE_URL for a different server address. The model needs to follow structured-output instructions.

Where does my code go?

Selected diff and repository content goes to your chosen provider. A local endpoint keeps that model call local. Reports stay on your computer or in your GitHub Actions outputs. Rival does not receive your code.

Does a low score mean it is safe to merge?

No. The score prioritizes review and shows its reasons. Brief reads code but never executes it or runs tests. Repository context can reveal more callers; it is not an exhaustive dependency audit.

What does it cost after buying?

Brief is a one-time purchase. The skill uses your existing coding agent’s allowance. The direct analyzer uses your chosen model account or local server; any cloud charges come from that provider.

Can it run on every pull request?

Yes. The included GitHub Actions workflow produces a run summary and downloadable HTML, SVG, and JSON reports. Forks and drafts are skipped. A local model needs a runner with access to its server.

What if it doesn’t work on my repo?

If Brief won’t run on your repository, or its report is plainly wrong, email support@rival.tips within 14 days of purchase with the output. We’ll help fix it. If we can’t, or you’d rather not wait, you get a full refund and your order link closes. Changing your mind about a working tool isn’t covered.

What am I buying?

The working CLI, installable agent skill, local viewer, automation workflow, examples, and editable MIT-licensed source. Keep and modify the downloaded version. The example reports need no API key.