Request audit

Send the agent PR reviewers keep questioning.

The first fit check is simple: one repo, one coding agent, one repeated PR pattern, and the review question that keeps coming back. Do not send source code, secrets, credentials, regulated data, or production access in the first email.

Price and scope: $750 for one repository and one primary coding-agent workflow. The audit returns files and a memo your team reviews before adopting.

Getting Started

Pick the repeated PR pattern

Choose the agent-created change that already raises review questions about scope, approval, commands, or evidence.

Send the context

Use the email template below. Do not send source code, secrets, credentials, regulated data, or production access in the first email.

Confirm fit and authorization

The first reply confirms timing, who can authorize the audit, and how the repository can be inspected locally by someone allowed to review it.

Review the returned packet

The audit returns repo artifacts, local check output, preflight notes, and a memo your team reviews before adopting any generated guidance.

Email Intake

Use the template so the first reply can confirm fit, timing, and authorization instead of starting from vague context.

Draft the request email

Direct email: [email protected]

What To Send

Subject: ContractForge audit for {company}

Company: Website: Repo name or public URL, only if safe: Coding agent used today: Agent-created PR pattern to audit: Where agent rules live today: Commands reviewers expect: Review question that keeps repeating: Who can approve the $750 audit: Timing:

Good Fit Examples

Package behavior change

Pattern: an agent edits matching logic, snapshots, or package metadata.

Question: what evidence should reviewers expect before accepting the PR?

API surface change

Pattern: an agent edits routes, response shapes, type definitions, or public API behavior.

Question: which tests, approval points, and skipped checks should be written down?

Backend-sensitive change

Pattern: an agent touches locking, hashing, migrations, billing, auth, or deployment-adjacent files.

Question: when should the agent stop, ask, or return a risk note?

Fit Check