Fixed-scope review
Codebase Autopsy
You inherited a system and need to know what you're holding
- Price
- From $4,000
- Turnaround
- Agreed before work starts
- Availability
- Taking work
When this is the right call
- You acquired a company and the software came with it
- A contractor or agency handed over a repository and disappeared
- Your team says it needs a rewrite and you have no way to check
- Nobody currently employed can explain how the system works end to end
- You are about to spend six figures on a platform you've never inspected
What you get
- The Salvage Report — architecture, risk, debt and unknowns, in writing
- Every finding ranked critical / high / medium / low, with reasoning
- A fix / defer / replace / do-not-touch call on each component
- Effort estimates against the ranked plan
- A 60-minute presentation to whoever is making the decision
What this isn't Not a code-style audit and not a linter report. Nobody needs to pay for an opinion about tabs. This is about what will hurt you, and when.
Every inherited system has three layers: what the documentation says, what the team believes, and what the code actually does. They are rarely the same document, and the gaps are where the money goes.
An autopsy closes the gaps and puts them in writing.
An autopsy is read-only. Salvage does not ask for write access to run one, and does not need it — everything in the report comes from reading the system, not from changing it.
What the report covers
Architecture. What exists, how the pieces actually talk to each other, and which dependencies are load-bearing. Drawn from the code, not from a diagram someone made in 2021.
Risk. Security, reliability, scalability, deployment, and data. Ranked by what would happen if it failed on a Tuesday afternoon, not by severity in the abstract.
Technical debt. Sorted into critical, high, medium and low — where critical means it is going to cost you something specific, and low means you can live with it for years and probably should.
Unknowns. The honest section, and usually the most valuable one. Undocumented behaviour, mystery dependencies, services still running that nobody can account for. Every inherited system has these. Most reports pretend otherwise.
The plan. Fix now, fix later, don’t touch, replace — with effort behind each call so it can be turned into a budget.
A full worked example of all of this — on a synthetic system, with the findings that actually recur — is published at the sample report. It is the whole deliverable at real length, so you can judge the work before buying any.
On rewrites
Roughly every team that inherits a difficult codebase asks for one. Sometimes that is the right answer. It is not the answer you should accept from anyone who has not yet read the system, including us — which is why the assessment comes first and is priced so you can afford to get it before committing.
The starting price covers a single application. Several services, a large legacy estate, or a report being used in an acquisition is quoted up front, before any work begins. The delivery date is agreed at the same time and put in writing with the scope — an autopsy is worth having on a date that holds, not on a date that was guessed to win the work.
A rewrite recommendation from someone who has spent a week in the code is a finding. The same recommendation on day one is a sales pitch.
Start with Codebase Autopsy
Describe the situation in whatever detail you have. If a different service is the better fit, you'll be told that before anyone talks about money.
First reply usually within a few business days, from a person