Job licence-and-snippet-scan-of-ai-written-code-with-a-list-for-your-lawyer · revised 12 October 2026
Scan AI-written code for licence and copied-snippet matches, with a list for your lawyer
A named scanner checks one repository's dependencies for licences and its AI-written code for matches to public code. You get a list for your lawyer. We give no legal opinion and no clearance.
You might be seeing
- A questionnaire asks about copyleft licences and AI-written code
- Nobody has listed the licences of the packages you ship
No passwords, keys, card details or admin invites needed to start.
What usually happened
AI assistants can reproduce licensed code, and dependencies bring their own licence terms, some of which place duties on what you ship. Whether any of it matters is a legal question, but a lawyer can only decide with a complete list. Nobody can honestly say the code is clear of all matches.
Who it’s for: Founder or engineering lead whose team uses AI coding assistants and who is asked by a customer, an investor or a buyer where the code came from.
Usually starts when: A due-diligence questionnaire asks about open-source licences and AI-written code, or you added a dependency and do not know its licence.
The result: You receive three lists for your lawyer, produced with the scanners named in the agreement: each inventoried dependency with its declared or reported licence, copyleft or unknown flag; each snippet match the tool reports above the agreed length, with its source link and reported licence; and files or file types the tool could not judge. The note states the scanner coverage and omissions; we give no legal conclusion or clearance.
Check whether this job fits
Answer these without sending files, tickets, code or logins. Nothing is submitted unless you choose to contact us.
What you get
- The dependency licence list
- The snippet-match list with source links
- The could-not-judge list
- The scanner names, versions, settings and commands
- A one-page note on what the lists cannot show
Included
- One repository at one named commit, read-only, with its dependency manifests
- A dependency licence list from a named licence scanner, with copyleft and unknown licences flagged
- A snippet-match run from a named tool over your own code, at a match length agreed in writing, with the source link and licence the tool reports
- A list of files and file types the scanners could not judge, such as generated, minified or binary files
- The commands and settings, so your team can re-run every list
Not included
- A legal opinion, a clearance, or a statement that code is free of infringement
- Removing or rewriting matched code
- A warranty or indemnity about the code or any dependency
- Judging whether a match is fair use, trivial or a real copy: your lawyer decides
- Scanning code you do not own or may not share
- Compliance with any specific licence, such as notices or source offers
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
Every package in the agreed manifest-and-lockfile inventory appears exactly once in the licence list, keyed by name and resolved version, with a declared or reported licence or an unknown flag. A script compares the complete key sets; one missing, duplicate or extra package fails the test.
Evidence: The complete inventory and licence-list key sets and the script output.
All 5 copyleft dependencies planted by your lead are flagged (5 of 5, because this checks a deterministic list). A missed planted copyleft dependency fails the whole test.
Evidence: The planted-dependency table.
At least 12 of the 15 planted snippets appear with source links, including all 5 your lead keeps from us until the run. One missed held-back snippet or false source link fails this smoke test; the count describes only these planted snippets and is not a detection-rate claim.
Evidence: The 15-snippet table, the 5 held-back marks and the checked source links.
A script confirms that 100% of the listed matches exist at the named commit with the lines shown.
Evidence: The citation check log.
Every list carries the scanner name, version, settings and the command to re-run it, and your team re-runs one list and gets the same rows.
Evidence: The commands and your re-run output.
Sign-off. You re-run the acceptance checks with the steps we hand over, read the results and sign off in writing. Payment follows sign-off.
If it fails. If an agreed check fails, you do not pay for this scope. We hand over what we found and agree whether to stop or re-quote; no surprise work.
When it fits, and when we stop
It fits when
- A repository you may share through a company-controlled route
- A lawyer or adviser who will read the lists and decide what matters
- A scanner that your team can run, or agreement that we run one on a copy
- 15 snippets of public licensed code and 5 copyleft dependencies planted in a copy; your lead keeps 5 snippets and all 5 dependencies from us until the test run
We stop and tell you if
- You want a clearance or a statement that no code is copied
- The code is under a contract that forbids sharing
- No lawyer or adviser will read the lists
- You want us to rewrite matched code
What could go wrong
The scan reads and reports, so there is nothing to undo. We delete our copies of your code when the job ends and the written terms say when.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| The lists are read as a clearance. | The note says in plain words that no scanner can show code is free of copied material, and the report lists what could not be judged. |
| A match is missed. | The report states how many of the 15 planted snippets were found. A scan finds only what its database and match length can find. |
| A match is wrong or trivial. | Each match shows the length, the source link and the licence as the tool reports it, so your lawyer can judge it. We make no finding of infringement. |
| The scanner's licence data is wrong. | Every licence is labelled as declared or as reported by the tool, with the file it came from. |
A separate automated reviewer checks the work and the test results before hand-over. Your lawyer or adviser reads the lists and decides what, if anything, to change. No human supervisor from us is included unless your agreement names one, and at launch the work is largely automated.
How we deliver
We arrange the work and independent review, then show you the result against the agreed checks. You keep authority over your systems.
- Agree commit, scanners, match length, adviser and terms in writing
- Receive code and manifests; your lead holds back 5 snippets and 5 dependencies
- Run the licence scanner and the snippet tool on a copy
- Check every listed match and licence against the tool's output
- Independent automated review of the lists and the note
- Send the lists, commands and note to you and your lawyer
This is a one-off job, not emergency cover or a subscription. We confirm eligibility, price, start window and delivery date in writing before you accept, and work starts only after the agreed inputs are in place. An enquiry creates no charge or booking.
Need to keep it working?
Discuss the monthly scan of new code once the first lists have been read.
Ongoing work is separately scoped and quoted: no monitoring, response-time guarantee or automatic subscription is included in this job.
Mention the responsibility you need in your enquiry.
What you can check
This is a new service. We have not delivered this job for a client yet.
Other ways to get this done
- Run an open-source licence scanner in your own pipeline and read its output yourself.
- Ask your lawyer to review your dependency list and your AI-use policy.
Questions
Will you tell me my code is clear?
No. No scanner can show that. You get lists and a note on what they cannot show.
Do you remove matched code?
No. Your lawyer and your team decide what to change.
Does anyone indemnify me?
No. We give no warranty and pass on no indemnity from any provider.
Who owns AI-written code?
That is a legal question and it differs by country. We give no view on it.
Does a pass mean it will be right next month?
No. A pass describes items like the test sample on the day of the test. It is not a promise about future items, and it does not replace your own checking.
Send an enquiry
Send us
- Languages and rough size of the repository
- What prompted the scan
- Which scanner you already use, if any
- No logins, keys, customer data or payment details in the first enquiry
Later, once you agree
- Read-only code at one commit through a company-controlled route
- The name of the lawyer or adviser
- A copy with 15 planted snippets and 5 planted copyleft dependencies; your lead keeps 5 snippets and all 5 dependencies from us until the test run
You keep the repository. We read one commit through a bot account with the least access that works, read-only first, and run the scanners on a copy in a sandbox with no secrets. Code is treated as data, and the agreement names any model provider, its country, retention and training status before anything is shared.
Email fallback: open your mail app
If website submission is unavailable, review and send the fallback email yourself. An email fallback is not a website receipt. Or write to hello@syntheticindustry.ai with “licence-and-snippet-scan-of-ai-written-code-with-a-list-for-your-lawyer” as the subject.