Software & Architecture.
Repository history, dependencies, database design, integrations, tests, maintainability, and places where change is unusually dangerous.
Software Project Rescue
When important software is late or unstable, “add developers” and “start over” are common answers. Both can be expensive mistakes.
The assessment is led personally by Codepoet’s founder and CEO. It starts with the existing system, identifies the parts worth keeping, and lays out the most defensible path forward for the current team and budget.
Standard assessment
$6,500 7–10 business days after access is completeDeadlines have slipped more than once and nobody can give you a dependable new date.
The team is busy, but finished work is hard to see.
Estimates and invoices are growing faster than the product.
A vendor or one key developer holds most of the system knowledge.
Releases are risky, manual, or regularly rolled back.
You inherited the software through an acquisition, departure, or failed handoff.
An application built quickly with AI now needs to handle real customers, data, and operational risk.
Evidence, not assumptions
A clean repository can still hide a broken release process. A busy issue tracker can still hide a team that is not shipping.
The assessment compares code, history, environments, and interviews before it reaches a conclusion.
Four evidence areas
Repository history, dependencies, database design, integrations, tests, maintainability, and places where change is unusually dangerous.
Environments, infrastructure, release paths, error visibility, backups, and the operational evidence needed to understand risk.
Authentication, authorization, secrets, sensitive data, and obvious security risk. This is not a penetration test or compliance certification.
Issue history, estimates, pull requests, meetings, roles, decision rights, delivery cadence, and team or vendor accountability.
The deliverable
This is not a code-quality score or a stack of scanner output. It is a decision document that connects evidence to risk, options, cost, and sequence.
Software Rescue Assessment
Executive verdict
Prioritized risk register
Decision matrix
30/60/90-day plan
Enough information to plan and compare options without pretending certainty that does not exist yet.
The decision-makers walk through the report, challenge the findings, and get the answers needed to act.
Because this is a new packaged offer, Codepoet does not present past development projects as completed Rescue Assessments.
Read the Fictional Sample AssessmentFrom evidence to action
Each stage narrows uncertainty. The work moves from establishing facts to giving the owner a sequence that can be funded and defended.
Establish the system, access, people, history, and business constraints actually in scope.
Compare code, issue history, environments, and interviews instead of accepting one account as complete.
Separate urgent business risk from ordinary technical debt and rank findings by effect, not drama.
Give the owner a defensible sequence: act now, plan next, monitor, or deliberately leave alone.
The engagement boundary
The review window begins when the agreed access is complete. Missing credentials, unavailable repositories, or absent stakeholders pause the clock rather than consuming the review.
Included in the standard assessment
Access boundary
Codepoet normally needs the repository, issue tracker, pull or merge request history, documentation, database schema, development environment, and relevant cloud or deployment information.
Codepoet will sign a reasonable NDA. Production credentials are not a standard requirement.
Your report. Your decision.
Implementation is proposed separately only after the findings are understood.
Prototype meets production
Software built with Claude, ChatGPT, Cursor, Lovable, Replit, or another AI-assisted tool can move from idea to working product remarkably quickly. The harder questions arrive with real users, sensitive data, integrations, billing, and multiple people changing the code.
The point is not to criticize how the software started. It is to determine what it needs next.
Production gates
One bounded application and up to three related repositories.
Several applications, broader infrastructure, regulated data, additional teams, active incident response, or unusually difficult access.
Scope and price are agreed before work begins. A standard assessment will not quietly become open-ended.Common questions
The assessment and any implementation proposal are separate decisions. The report explains the evidence behind each practical option and is written so your current team or another qualified provider can act on it. Codepoet only scopes follow-up work if you ask.
Yes. A rescue does not automatically mean replacing the team. Good developers can struggle inside a project with unclear priorities, weak ownership, or a delivery process that works against them.
That is common. Repository history, issue records, infrastructure, interviews, and the running system can still provide substantial evidence. Missing documentation is identified as a risk rather than guessed around.
Usually not. Codepoet prefers repositories, documentation, schema information, a development environment, and read-only operational evidence. If production access becomes necessary, Codepoet will explain why before requesting it.
Not as part of the standard assessment. Separating diagnosis from implementation protects the system and keeps the findings independent. Emergency work can be authorized separately.
The founder’s current hands-on work is primarily PHP, Laravel, React, Tailwind, and Linux, with decades of earlier C# and .NET experience. Many assessment questions cross language boundaries. If specialist review is needed, Codepoet will say so before accepting the work.
The first call is 30 minutes. We will discuss what is happening, how urgent it is, and what access a useful review would require.
Send a Message