Applied AI-Powered Bioinformatics Tools

Chain-of-Custody Case Tracker

Last reviewed

Beginner

What you'll learn

~45 min
  • Ask an AI tool, in plan mode, for a program that reads a training mock case's typed custody records (the case's status, its item log and the lines of each item's custody form, one line per hand-over) and the laboratory's numbered custody rules, lays out each item's custody in line order, and lists what needs a look
  • Point it at the laboratory's rules file, and read its plan for what matters: every rule in it by number, items matched both ways, line order kept, every blank marked, each mark naming its rule, and only the words you gave it
  • Check its list against findings worked out for the practice files
  • Say what a mark and the absence of one can tell you, and who decides what a mark means for the record

The story, and what you will end up with

It is Thursday 12 November 2026 at a university laboratory that receives items and samples from field recovery projects and outside organisations. You are the laboratory’s evidence custodian: you run its evidence store and keep its custody records.

Each autumn the laboratory runs a mock case to train new staff in its custody procedures. This year’s, TRN-07, ran from 2 to 6 November. Ten items that stand in for real ones, every package and form marked TRAINING, NOT EVIDENCE, went from a trainee acting as field lead to a courier, from the courier to your deputy at the laboratory, into the evidence store, and most went out to two trainee analysts for sampling and back. On 5 November the trainees collected four more from a fourth training plot, which reached the laboratory the next morning. The exercise’s timetable closed the case at 17:00 on 6 November, and you entered the closing in the case register; reconciling its record is the debrief’s work. In a real case, your procedures say what is reconciled before a case is closed. Each item has its own custody form: one numbered line for each hand-over, with who released it and when, who received it and when, the purpose, both signatures, and, on a line where the evidence store releases or receives the item, the custodian who signed for the store. A colleague has typed the case’s entry in the case register, the item log and every form’s lines into three files. The laboratory’s custody rules, numbered, are in a fourth. The debrief is tomorrow morning.

Before you build anything, you have done what comes first. When the courier’s packages arrived, your deputy checked each seal and the item count against the courier’s manifest with the courier present. You have the signed forms in hand, with the courier’s tracking record and the evidence store’s own access log, which shows who went into the store and when; the practice files hold none of these. This morning you walked the store: the bags for T7-01 to T7-10 and T7-12 to T7-15 are on the shelf with their seals as recorded, and so is a bag marked T7-11, which is not in the item log; you have set it apart and noted it, as this practice laboratory’s procedures say. Nobody has yet checked the typed lines against the signed forms, so for now the typed files are an unchecked copy, and that check is still to be done whatever the list shows.

What you want before the debrief is every item’s custody laid out line by line, and every place where the written record does not account for an item marked, with the rule and the reason. Today that means reading 61 typed lines by eye against fourteen items and ten rules.

By the end of this lesson you will have a small program on your own computer that reads the files, lays out each item’s hand-overs in the order of its form’s lines, and lists what needs a look and under which rule, on screen and as a file your spreadsheet program opens. It reads only what was typed. It never says custody is intact, broken, acceptable or compromised, and it never decides what a mark means. You build it by asking your AI tool in plain English, and you check it against findings this lesson gives you.

🔍Words used in this lessonclick to expand
  • Custody form: the paper form that goes with one item, one numbered line for each hand-over. In this practice laboratory the signed form is the record; the practice files are a typed copy of it.
  • Hand-over: one line of a custody form: the item released by one person or place and received by another.
  • Line number: the number of a hand-over on its item’s form. It gives the order the hand-overs were written in.
  • Item log: the list of items logged into the case, one row per item, with who logged each and when.
  • Case register: the laboratory’s list of its cases, with whether each is open or closed.
  • Store A: this practice laboratory’s evidence store, written in the forms as a holder in its own right. On a line where the store releases or receives an item, the form names the custodian who signed for the store, and the store’s signature is that custodian’s; the typed copy says only that a signature is there. Who went into the store, and when, is in the store’s own access log.
  • Rules file: this practice laboratory’s custody rules, numbered and in ordinary sentences. Your laboratory’s are in its own written procedures.
  • Extract: material taken from an item, such as by an analyst sampling it, that is then handled as an item of its own.
  • Last holder by the record: whoever received the item on its last typed line. It is what the record says, not proof of where the item is.
  • Needs a look: what the list says about anything the written record does not account for under your rules. No mark from this check means only that this check, on these files, found nothing under those rules.
  • Checksum: a short code worked out from a file’s exact contents. If the file changes at all, the code changes, so it shows which version of a file a list was made from.

Before you start

First, if your AI tool is still open, leave it: type /exit and press Enter (if you use Antigravity, press Ctrl+D with nothing typed instead), and wait until you see the $ or % prompt. The next line goes there, in the terminal, not in the AI tool. It makes a folder for this lesson beside your practice folder and goes into it:

Terminal window
mkdir -p ~/bio-workbench/13-custody-check && cd ~/bio-workbench/13-custody-check

Start your AI tool with the start line you kept from the setup lesson, the one with approval prompts off. Then ask it to set the folder up:

Set this folder up so we can undo any change you make, download https://learn.intellimetrics.net/starters/biotech/13/custody-records.zip and unpack it into a data folder, and save a checkpoint.

Look for it saying three things: the folder is set up with Git (the program that saves snapshots, from setup), the four practice files are in the data folder (trn07-case.csv, trn07-items.csv, trn07-handovers.csv and trn07-custody-rules.txt), and a first snapshot, your checkpoint, is saved. Later you can ask whether the files have changed since then. If it does not mention one of them, ask: “Did you save a checkpoint with the files in it?” It may also keep the zip file in the data folder, which is fine; if it says it kept one somewhere outside this folder, ask it to delete that copy. If the download does not work, download custody-records.zip in your browser, check where it went (the browser’s list of downloads shows the folder), and tell your AI tool, for example: “custody-records.zip is in my Downloads folder. Unpack that one file into a data folder here, and save a checkpoint.” In Windows PowerShell or a desktop app, “my Downloads folder” works as is; only say “my Windows Downloads folder” if you use Ubuntu (WSL) on Windows, which keeps its own Downloads folder apart from Windows.

The box below says what is in the practice files; later steps refer to them.

ℹWhat is in custody-records.zipclick to expand

Three files, each of which opens in a spreadsheet program, and a rules file that opens in any text editor. Every row of the three begins with the marking TRAINING, NOT EVIDENCE and the case number TRN-07, and the rules file begins and ends with the marking.

  • trn07-case.csv: the case’s entry in the case register. Columns: marking, case, description, opened_at, status, closed_at, closed_by. One row: TRN-07, opened 2 November at 08:00, closed 6 November at 17:00 by A. Novak.
  • trn07-items.csv: the item log. Columns: marking, case, item, description, packaging, collected_at, logged_by, logged_at. Fourteen items, all logged by K. Varga: T7-01 to T7-10 on 2 November, T7-12 to T7-15 on 5 November.
  • trn07-handovers.csv: the typed lines of every item’s custody form, 61 rows, grouped by item: five lines each for T7-01 to T7-07 and T7-15, seven for T7-09, four for T7-13, three each for T7-10, T7-12 and T7-14, one for T7-11 and none for T7-08. Columns: marking, case, item, line, released_by, released_at, received_by, received_at, purpose, released_signed, received_signed, store_custodian. In the two signature columns, yes means the typist saw a signature on the form; a blank means none was typed. The store_custodian column is filled in only on lines where Store A releases or receives the item.
  • trn07-custody-rules.txt: the practice laboratory’s custody rules, below.
ItemDescriptionCollected at
T7-01, T7-02soil sampletraining plot 1
T7-03, T7-04, T7-05bone fragment (as submitted)training plot 2
T7-06textile fragmenttraining plot 3
T7-07metal objecttraining plot 3
T7-08soil sampletraining plot 3
T7-09textile fragmenttraining plot 1
T7-10soil sampletraining plot 2
T7-12, T7-15soil sampletraining plot 4
T7-13textile fragmenttraining plot 4
T7-14metal objecttraining plot 4

Every item is in a sealed bag. “As submitted” means the description is the one the item was logged with.

Name in the formsWho or what
A. Novakyou, the evidence custodian
K. Vargatrainee acting as field lead
D. Oseitrainee acting as courier
T. Marshyour deputy custodian
L. Ferrand, S. Brandttrainee analysts
Store Athe laboratory’s evidence store

An item’s usual lines: K. Varga to D. Osei at 16:10, for transport to the laboratory (2 November for the first ten items, 5 November for the last four); D. Osei to T. Marsh at 09:20 the next morning, receipt at the laboratory; T. Marsh to Store A, storage, five minutes from release to receipt (the walk to the store); then, for most items, Store A to an analyst for sampling, and back to Store A. Times are local time, written year-month-day and 24-hour, such as 2026-11-03 09:20.

The rules, as the rules file has them:

  1. Every item in the item log has at least one hand-over line, and every item on a hand-over line is in the item log. Items are matched by case and item number exactly as written.
  2. Each item’s line numbers start at 1 and run in order with no gaps and no repeats.
  3. For an item in the item log, its first line is released by whoever logged the item, and not before the time it was logged.
  4. Whoever receives an item on one line releases it on the next line, with the name written the same way.
  5. No line is dated before the line above it: each release is at or after the receipt on the line above.
  6. An item has one holder at a time: on each line, the receipt is not dated before the release.
  7. On each line, the receipt is no more than 15 minutes after the release.
  8. Every line has who released the item and when, who received it and when, the purpose, and both signatures. Any of these left blank, and any time that cannot be read, needs a look. Nothing is filled in or guessed.
  9. A line on which Store A releases or receives the item names, as store custodian, the custodian who signed for the store, and the store’s side of the line is signed.
  10. When the case is closed, an item whose last line leaves it with a person, not in Store A, needs a look.

Everything in these files is made up for this course: the mock case, every item, person, place, time and rule. Each lesson in this track has its own practice files, with their own people and dates.

Switch to plan mode

Your AI tool is already open. Switch to plan mode now: Shift+Tab until it says plan, or /plan (the keys for each tool). In plan mode it usually reads your files, asks its questions and shows you a plan before it builds anything; the order can vary.

The conversation

Each message below is an example. Say it your own way.

Say what you want

Say what you do, which files to use, and what you want from them:

I'm our lab's evidence custodian. Build a program that lays out each item's custody, line by line, from the records in data, and lists what needs a look under our rules. Ask me what you need first.

Your AI tool usually reads the files and comes back with questions, starting with your rules. It may already point out some of what it sees in the files. Leave its questions for now: send the three messages below first, then answer them. Your AI tool may wait for your answers or go straight on to propose a plan; either way, do not approve a plan until the three messages are sent.

Point it at your rules

These are this practice laboratory’s own rules, in its rules file. Your laboratory’s are in its own written procedures; later in this lesson you write them into a file of your own. Even if your AI tool has already read the rules file, tell it these are the rules to follow:

Our facility's custody rules are in trn07-custody-rules.txt in data. Mark whatever breaks one of them, and say beside each mark which rule it breaks.

Wait for its reply, then say how the list is worded:

In the list, say only "needs a look" or "no mark from this check", and "last holder by the record", never intact, broken, acceptable or compromised. Head each list with the date, the files read with checksums, and these rules.

Wait for its reply, then say what the program must do with the files:

Make it a program I can run again on each case's files. Count the rows read from each file, list any row it can't read with the reason, leave my files alone, and never write over an earlier list.

The 15 minutes in rule 7 is this practice laboratory’s own figure, chosen so the walk from the receiving desk to the store does not need a look. Your procedures set yours, if they set one. The words “needs a look” and “no mark from this check” are there because the list has checked typed files and nothing else: it cannot say whether custody held. The header says which files, in which version, under which rules, because a list like this may itself be asked for later.

What a custody record must hold is for your procedures to say. One published list is in a handbook for evidence handlers from the US National Institute of Standards and Technology, which says that “Every transfer of evidence between individuals and storage locations must be documented.” and that “Chain-of-custody documentation should include the following:” description of the evidence; unique case identifier (e.g., case number); where the evidence was collected; where the evidence was stored; who was in possession of the evidence and for what purpose; what was done to the evidence (e.g., analysis or re-packaging); date and time information (NIST, The Biological Evidence Preservation Handbook, NIST IR 7928, April 2013, page 25, read 12 November 2026). The practice files carry each of these, in the item log and the purpose column. Your laboratory’s procedures, and its accreditation if it has one, say what yours must hold.

Answer its questions

Answer each in a sentence. These are questions the tools often ask, with one way to answer each. Yours may ask them in a different order or show a list of choices:

If it asksYou might say
Is there a list of who may sign for Store A?Not in these files. Check only what rule 9 says.
A blank time stops some checks. Say they can’t be checked?Yes. A check it cannot make is never a pass.
Compare T7-09’s line 5 with line 7, across the missing line?Yes, compare across it; the missing line itself needs a look under rule 2.
Mark a hand-over dated after the case was closed?No rule here covers it. List any apart, outside the rules, for me to look at.
Only this case, or any case’s files in the same layout?Any case whose files have the same layout.
On screen, a spreadsheet file, or a page?On screen, and a file my spreadsheet program opens, each with the header.
What should it be written in?Use whatever needs the least installed and works with no internet.

If it asks several questions at once, answer the ones about your rules and how hand-overs are judged from your laboratory’s own procedures, a sentence each, and accept its suggestions only about how the list looks. If you cannot tell which kind a question is, treat it as a rule. If it asks something not listed here, answer it the same way, in a sentence.

Read the plan

Next it usually shows you a plan: what it will build and how it will check it. Read it for six things:

  1. Every rule in the rules file is in the plan, by its number, and every mark will say which rule it is for. Look for all ten, and in particular for the first line checked against who logged the item and when (rule 3), a receipt dated before its release (rule 6), the custodian named on every Store A line (rule 9), and an item left with a person in a closed case (rule 10).
  2. Items are matched both ways between the item log and the hand-over lines, by case and item number exactly as written, and every row of every file is counted, with any row it could not read listed and its reason.
  3. Each item’s lines are read in line order, never re-sorted by time, and each item shows its last holder by the record.
  4. Every blank the rules name needs a look, signatures included, and is never filled in or guessed.
  5. Everything that needs a look gives its rule and its reason, the list uses only the words you gave it, and each list is headed with the date, the files read with their checksums, and your rules.
  6. It is a program you can run again on the next case’s files, your files are read and left as they are, and an earlier list is never written over.

A first plan often leaves one of these out, even when you have asked for it. Say so in a sentence, in your own words, for example:

The plan doesn't say how it checks rule 9. Every line where Store A releases or receives an item needs a named custodian and the store's signature.

When the plan has all six, approve it: if your AI tool offers a choice to go ahead, pick it; otherwise say “Yes, go ahead and build it.” In Codex, press Shift+Tab first to leave plan mode. In Copilot, saying it brings up a choice to go ahead; pick that. Let it build. It may run what it built and tell you how to run it yourself; either way, the next step asks it to. When it has finished, ask it: “Did you install anything, and where?” Nothing, or something inside this lesson’s folder, is fine for these lessons. If it names anywhere else on your computer, such as the zip file it downloaded, ask it what that is for, and to remove it if you do not want it there.

Look at what it built

Ask it to run what it built:

Run it on the files in data and show me the list.

Your list will be worded and laid out in its own way, and some details may be only in the spreadsheet file, not on screen. Each mark should name its rule by number. Check it for these:

  • The header gives the date, the files read with a checksum each, the rules, and the rows read: 1 from the case register, 14 from the item log and 61 from the hand-overs, none it could not read.
  • T7-08 needs a look under rule 1: it is in the item log and has no hand-over line.
  • T7-11 needs a look under rule 1: it has a hand-over line, L. Ferrand to Store A on 5 November, and is not in the item log. Its purpose says it is an extract from T7-05; the files say nothing more about it.
  • T7-09 needs a look under rule 2: its lines go from 5 to 7, with no line 6. The lines either side still follow on: line 5 ends in Store A and line 7 starts there.
  • T7-12 needs a look under rule 3: K. Varga released it on line 1 at 16:10 on 5 November, and the item log says it was logged at 16:25.
  • T7-03 needs a look under rule 4: line 4 hands it to L. Ferrand, and line 5 has S. Brandt releasing it.
  • T7-05 needs a look under rule 5: line 5, the return to Store A, is dated 09:30 on 5 November, before line 4, the sign-out for sampling, at 10:00.
  • T7-15 needs a look under rule 6: on line 4, Store A to L. Ferrand, the receipt is at 11:00 on 6 November, five minutes before the release at 11:05.
  • T7-04 needs a look under rule 7: on line 2, D. Osei released it at 07:15 on 3 November and T. Marsh received it at 09:20, 2 hours 5 minutes later.
  • T7-06 needs a look under rule 8: line 4, Store A to S. Brandt, has no receiving signature.
  • T7-07 needs a look under rule 8: line 5, S. Brandt to Store A, has no receipt time. Some lists also mark rules 6 and 7 on that line, saying the time checks could not be made; that is the same blank, not a second problem.
  • T7-14 needs a look under rule 9: line 3, T. Marsh to Store A, names no custodian for the store.
  • T7-13 needs a look under rule 10: the case is closed, and its last line, line 4 on 6 November, leaves it with S. Brandt.
  • T7-01, T7-02 and T7-10 have no mark from this check. Some lists name them; others list only what needs a look.
  • The last holder by the record is Store A for every item with a line, T7-11 included, except T7-13, whose last holder by the record is S. Brandt. T7-08 has none, or the list says it cannot be told.
  • No line calls custody intact, broken, acceptable or compromised.
  • When you ask “Have my record files changed since the checkpoint?”, it compares them with the checkpoint and says no.

If something on the list is not true, tell your AI tool in a sentence what you see and what you expected, for example: “T7-14 shows no mark. Rule 9 says a line into Store A names the custodian who signed for the store, and line 3 names none.”

If a line or a reply says custody is intact, broken, acceptable or compromised, or says what a mark means, say: “Take that off. The list says only needs a look or no mark from this check, never what a mark means.”

Notice what no mark means

Look at T7-04 on your list. Nothing in the files says where the item was in the two hours between D. Osei’s release and T. Marsh’s receipt. Now suppose that someone, meaning well, tidied the typed record so the times agree. Try it on a copy:

Make a copy of data/trn07-handovers.csv in this folder. In the copy only, change T7-04's line 2 receipt time to 07:15. Then run the program on the copy and show me T7-04 beside the original.

Before you read on: what will the list say about T7-04 now?

No mark from this check. The copy meets every rule the program checks, T7-04 drops off the list, and the list from the copy looks just like a list made from a record with no gap in it. Nothing in the copy says it was changed. The header gives it away only because it prints a different checksum for the hand-over file, and only if someone compares it with the original. The signed form still says 09:20. Changing a typed time changes nothing about what happened, and in this practice laboratory the signed form is still the record.

The program reads what was typed and nothing else. It cannot tell a tidied time from a true one, a time written in error from a hand-over that really happened late, or a line typed wrongly from a line that was never written. So a mark is a question, not an answer, and every mark is settled against the signed form first, because the typed files are a copy of it.

In this practice laboratory, an original line is never overwritten or tidied: a correction is a new line, dated and initialled when it is made, that says what it corrects. (That is this practice laboratory’s own rule; your procedures say how yours are made.) When you have looked, remove the copy: “Delete that copy; I’m done with it.”

Then the marks themselves. Each one is a question for the people and records the program never reads:

  • T7-04: where was the package between 07:15 and 09:20? Was it left somewhere before T. Marsh received it, such as at reception or in a drop locker? If so, a log kept for that place, and the courier’s tracking record, may show where it was; your procedures say whether such a stop belongs on the form as a line of its own.
  • T7-05: was the return written as 09:30 when it was 10:30, or written up later? Ask L. Ferrand, and look at the store’s access log for that morning.
  • T7-15: one of the two times on line 4 is wrong, on the form or in the typing. Compare the typed line with the signed form and the store’s access log for 6 November, and ask L. Ferrand, who received it. You signed for the store on that line, so your procedures may say that someone other than you looks at it.
  • T7-03: L. Ferrand received the item, and S. Brandt returned it. Was it passed between the two analysts with no line written, or is one name wrong on the form?
  • T7-12: the item log says 16:25, after the item left with the courier at 16:10; the other three items from that plot were logged before 16:10. Was it logged late, or is the time a slip? Compare the typed time with the signed item log, and ask K. Varga, who logged it.
  • T7-13: its bag is on the shelf, yet its last typed line leaves it with S. Brandt in a closed case. Is the return line on the signed form and not typed, or was it never written? Ask S. Brandt, and look at the store’s access log for 6 November.
  • T7-14: its bag is on the shelf, and the typed line says the store’s side was signed. Whose signature is it? Compare the typed line with the signed form and the store’s access log, and ask T. Marsh.
  • T7-08: it is on the shelf and in the item log, with no line typed. Was its form not typed, or never started? Look for its form among the signed forms, and ask the colleague who typed them.
  • T7-11: its bag is on the shelf. An extract made from T7-05 is an item of its own, and in this practice laboratory it is logged, with its parent, when it is made. The typed item log does not hold it; look for it in the signed item log, and ask L. Ferrand, named on its one typed line.
  • T7-09: is line 6 missing from the form, or was it not typed? Compare the typed lines with the signed form.
  • T7-06 and T7-07: is the signature or time missing from the form, or only from the typed copy?

In this mock case, what you find goes to tomorrow’s debrief. In a real case, who decides is set by your laboratory’s procedures; check which record holds that decision. The list and anything your AI tool drafts decide nothing.

A custody check marks where the typed record does not account for an item under the rules it was given; no mark is not proof that it does, and what each mark means, and whether the record can be relied on, are for the evidence custodian and whoever the laboratory’s procedures name to decide, from the signed originals and the people involved.

One real run

Here is part of what one AI tool said after the experiment above, with Claude Code on 12 November 2026. Yours will word it differently; check your list against the checks above, not against this.

So the change takes T7-04 off the list completely. Every other
item's marks in the copy's list are the same as in the original's.
The list heading also shows that a different handovers file was
read: its checksum is `2b3c40da…d38db9c6`, against the original's
`e121ae5d…202bebb`.

The practice file contradicts one part of this: the original hand-over file’s checksum ends 4e2bebb, not 202bebb. A shortened checksum can be shortened wrongly, so when you compare two lists’ headers, compare the checksums in full.

Here the copy kept the original’s file name, so only the checksum in the header shows which file the list was made from. Neither shows which version of a time is true; the signed form is the first record to compare, and it may not settle what happened.

The mock case and the records it talks about are made up for this course.

How real records differ

  • In this practice laboratory the typed copy is not the record. A yes in a signature column is the typist’s note that a signature is on the form; the signed form is the record, and every mark is settled against it. Check which record your own laboratory’s procedures treat as the record, a signed form or a signed entry in an evidence system, and who checks a typed copy against it, and when, before anything is read from it.
  • How are your hand-overs numbered? In these practice files a line number is the line on the item’s own form. Find out when and how your records’ numbers are given, what order an export of them comes in, and whether anyone can change them, before the program reads them as the order of hand-overs.
  • Is your courier’s record for a package or for each item? The practice files write each item’s courier lines separately. Check whether your records tie items to a sealed package with a manifest and a tracking number instead; if they do, the program would need to read both.
  • How do your records name the store? In these files Store A is written as a holder, and the form names the custodian who signed for it; the store’s access log is a separate record, and the program does not read it. Check how your records name the store and who signed for it, how entries to your store are recorded, and whether your procedures have them read beside the custody lines.
  • What is recorded when an item comes back from sampling? These files record no seal condition, no reseal and no note of what was used. Check what your procedures have recorded then, and how a reseal is marked.
  • What do your records hold that these do not? Check whether yours hold handwriting keyed in by someone else, illegible entries, struck-through corrections with initials, lines marked as late entries, times in more than one time zone, 12-hour and 24-hour times mixed, a year left off, or a person’s name written differently from form to form. Each of these needs a rule your procedures set, or it needs a look.
  • The list is itself a document about the case. Check how your laboratory’s procedures say a working document made from a case’s records is handled and kept, like the records it was made from.

Keep going

Small changes do not need plan mode. Say each in a sentence, then run it again:

  1. “Make a sheet for each item that prints its custody, line by line, with the list’s header.” A printed sheet from this program is not your laboratory’s custody form unless its procedures say so. Mark it the way they say working copies are marked.
  2. “Does the program name any item or person from these practice files? If it does, make it work those out from the files instead.” A program built in a conversation may keep your examples written into it, and on the next case’s files it would get them wrong.
  3. “Put the program’s version in each list’s header, so a list shows which version made it.” A list may be asked for long after the program has changed.
  4. “Save a checkpoint.” Your AI tool saves a snapshot you can return to.

When you want to check a real case’s custody records, use the finished program. It runs on your real files by itself, with your AI tool closed, so your AI tool never sees them. Nothing confidential or unpublished, and nothing about people or about animals under a protocol, goes to your AI tool or into this course folder. Records of a real case never go to your AI tool: not a form, not a row, not a name from them, not a screenshot of an error. Whether they may go anywhere else is set by your laboratory’s policy, its custody procedures and any agreement or order that covers the case. Your evidence system’s export, or the sheet you type the forms into, may not be shaped like the practice files; check its columns against theirs. Do these steps in order; the first five happen now, while your AI tool is open and this folder still holds only the practice files:

  1. Go through the data check, the questions in setup about what the data is, who decides how it is used, and which AI tool would see it. Go on only if the people who decide allow this use.

  2. Decide on a folder of its own for your real files, outside this course folder, for example ~/custody-records. Nothing has to be in it yet. If the folder is not there, ask your AI tool to make it: “Make an empty folder called custody-records in my home folder.”

  3. Ask your AI tool for blank files to fill in, and to show the program working on them:

    Make blank copies of the three record files in ~/custody-records, same columns, with one made-up item and its hand-overs, that my spreadsheet opens. Keep the program and its checkpoints here. Run the program on them and show me.

    Look for it running the program on the made-up item and showing its lines and its result. It may also put a copy of the rules file there, since the program reads one beside the records; the next step replaces it with yours. Your AI tool may warn that a spreadsheet program can rewrite times when it saves, such as 2026-11-10 09:30 becoming 11/10/2026 9:30, which the program would then mark as unreadable; ask it how to avoid that in your spreadsheet. If your evidence system’s export has other column names, tell your AI tool those names, with two rows you make up in the same shape, made-up items, names and times, never a real row, and ask it to read files like that.

  4. Have your AI tool make a rules file for your laboratory’s own custody rules, while only made-up data is there:

    Make a plain-text rules file in ~/custody-records, named the way the program looks for one, with one placeholder line. Tell me its name.

    A rules file holds rules, not records: no case, item or person from a real case goes in it. Your AI tool reads the rules file, so if your laboratory treats its procedures as confidential, ask whoever owns them before any of their wording goes to it. Then open that file in any text editor and replace the placeholder line with your laboratory’s own custody rules, from its written procedures, the way the practice file has them: numbered, one rule to a paragraph, in ordinary sentences. Save it under the same name, and tell your AI tool:

    Our rules are in that file now. Change the program to follow those rules instead of the practice ones, and update the blank files to match.

    If your procedures keep a list of who may sign for the store, ask for the program to read it from a file beside your records, and test it with two made-up names; your real list never goes to your AI tool. It may add columns your rules need, such as a witness and a witness signature, and say what it changed; if it asks how a rule fits, answer from your procedures, a sentence each. It may also say the program no longer runs on the practice files, because they do not fit your rules; that is expected. Wait for its reply, then:

    Make up one item for each of our rules that only that rule marks, and one with nothing wrong, and show me the list.

    Where your rules overlap, an item may be marked by more than one, and the tool should say every rule that marks it.

    Check that each made-up item needs a look under the rule it was made for, with that rule’s number and reason, and that the made-up item with nothing wrong has no mark. The program follows the rules it was built for, and your AI tool may build it to stop when the rules file no longer matches them. If your procedures change a rule, change your rules file, ask your AI tool to change the program to match while only made-up files are in your folder, and check it the same way.

  5. Ask your AI tool for the exact line:

    Write me the exact line to run this myself on the files in a folder called custody-records in my home folder, with full paths, saving everything it writes there, never in this folder.

    Use your own folder’s name. If it offers to copy your real files into this folder instead, say no. Copy the line it gives you into a note. It will look something like this one, though yours may differ:

    Terminal window
    python3 /home/dana/bio-workbench/13-custody-check/tools/custody_report.py /home/dana/custody-records --out /home/dana/custody-lists

    Read it left to right. python3 runs a Python program. Next comes the checking program, in this lesson’s folder; then the folder that holds a case’s files, which the program finds there by name, your rules file included; --out and the folder after it say where the list goes, a new file each run. Your AI tool may say the program will not write into the folder it reads, and send the list to a folder beside it, such as custody-lists, which keeps your records apart from what the program writes. The line depends on the computer: the one above came from a Linux computer, so use the line your own AI tool gives you. /home/dana is the home folder of someone whose user name is dana; your own user name goes where dana is, and on a Mac the folders begin /Users/ instead of /home/. The line runs the program where it is, so leave this lesson’s folder in place. Then ask your AI tool what the program needs in that folder: “What must the three record files and the rules file be named, and which columns must each record file keep?” Note its answer beside the line. If you are unsure about any part, ask your AI tool now, while it is still open.

    Then, before you leave your AI tool, try the line on the made-up files, because a real case’s records must never end up in the course folder: “Run that line on the made-up files in ~/custody-records. Where did the list land, and did running it change anything in this folder?” Look for it saying the new list is in the folder your line names and nothing in this lesson’s folder changed. If the list landed in this lesson’s folder, tell your AI tool so and ask it for the line again.

  6. Leave your AI tool the way you did at the start of this lesson.

  7. First remove the made-up items and the lists made from them; your rules file stays. Or have your AI tool do it before you leave it in step 6, while only made-up data is there. Then put a real case’s files in that folder under the names you noted in step 5, beside your rules file. As setup says, a tool started with approval prompts off can open other files your account can open, so from now on start your AI tool only in your course folder and never ask it about the custody-records folder. Paste the line from your note at the $ or % prompt and press Enter. The list appears in the folder your line names, not in the course folder. To come back to the lesson folder later, type this lesson’s folder line again, the line near the top of this page that begins with mkdir.

  8. Before you rely on the list for a new case, run it on a past case whose record you have already checked yourself, and compare every mark with what you found. If they differ, find out why first. And settle every mark against the signed forms, or whatever record your laboratory’s procedures name, not the typed copy or an export.

Try another case

Before the quiz, write two or three sentences for yourself. A month from now, on a witness stand, you are asked why an item’s form shows it received by one analyst at 14:10, and the next line shows a different analyst releasing it at 14:30. Your list marked that line. What do you say, and what do you look at before you go?

A good answer goes to the signed form first, to see whether the typed lines match it, then to the store’s access log and the two analysts for what happened between those lines. It says only what the records show and what was done about the mark when it was found, and it does not offer the list as an answer: the list showed where to look.

Key takeaways

  • A short ask, then your rules file. Say what you do and which files; point it at your laboratory’s numbered custody rules, written from its own procedures.
  • Read the plan for every rule by number: items matched both ways, line order kept, each receiver the next releaser, every blank marked, each mark naming its rule.
  • Check a new list against findings you know before you trust it on a case you do not.
  • No mark is not proof. A tidied record passes every check under the rules it was given; what a mark means, and whether the record can be relied on, are for the evidence custodian and whoever your laboratory’s procedures name. In this practice laboratory the signed form is the record.
  • Practice data stays practice data: real records go to the finished program, with your AI tool closed, after the data check.
KNOWLEDGE CHECK

Your list shows no mark for an item. Line 2 of its typed form says the courier released it at 07:15 and your deputy received it at 07:15. The signed form says received at 09:20. What does the list's result tell you?

What’s next

The next lesson is about laying a questioned sample’s allele calls beside a reference sample’s and saying only what the two rows hold.

intellimetrics · Arlington, VA
© 2026 intellimetrics.net
Search lessons