Real estate cyber security: follow a rental from application to lease, and test who reaches each record
A guide for the principal, licensee or property management lead of an Australian real estate agency.
It follows a rental through inspection bookings, applications with identity and bank documents, the lease, maintenance requests, contractor work orders, owner portals and changes to payment details, and comes with a rental-to-lease checklist of tests to run on your own systems with invented applicants, property managers, landlords and contractors.
On this page
Rental-to-lease checklist for agencies and property managers
Twenty-five rows in nine groups: I inspection enquiries and bookings, R rental applications with identity and bank documents, D declined and expired applications, L the lease and tenancy documents, M maintenance requests and photos, C contractor access, O owner portals, P changes to payment details, and V third-party systems. Run it only on systems the agency owns or controls, on a staging copy where one exists, with invented test roles: applicants AP1 (approved for test property X, then its tenant), AP2 (declined for X) and AP3 (tenant of test property Y); property managers PM1 (X) and PM2 (Y); landlords LL1 (owner of X) and LL2 (owner of Y); and contractors CT1 (assigned a job at X) and CT2 (assigned a job at Y).
How to use itHide the instructions
Every identity document and bank statement is made for the test and marked as a specimen, and every email address and phone number belongs to the agency. The expected results shown are examples: replace each with what the agency's written rule says before the first request. Five rows only read and submit nothing: I2, R1, D1 and P1 read a form, settings, a rule or a list, and V1 reads a vendor's terms and asks the vendor a question. The other twenty rows sign in, submit or send requests, and the last cell of each starts with the conditions the row runs under, then states how many requests it sends. Any link that opens a record without a sign-in is opened only for the invented records and has its token redacted from the evidence. A real person's data in any response ends the row: note what showed and where, go no further, and pass it to whoever handles privacy for the agency. No row runs against a vendor's system until the agency has read that vendor's terms and asked the vendor.
25 rows
Step and check
Data involved
I1. Inspection bookings: what a booking reference opens
- Data involved
- Names, phone numbers and email addresses of the people booked to inspect a property, and the times they chose. In the test, AP1 and AP2 booked for property X.
- Who should and should not reach it
- Should: each person for their own booking, and PM1 for the list of everyone booked. Should not: another person who booked, anyone who changes the reference in a link, or anyone signed out.
- Test case
- Book one inspection of property X as AP1 and one as AP2. Open AP1's confirmation link or booking page, change the booking reference or identifier to AP2's and request it. Then request the attendee list that PM1's view loads, once from AP1's session or link and once with no session (WSTG-ATHZ-04; ASVS v5.0.0-8.2.2).
- Expected result
- AP1 sees only AP1's booking. The changed reference and both attendee-list requests are refused, with the same response as a booking that does not exist.
- Evidence to keep
- Both confirmations as received with any link token redacted, each request and response, and the form of reference the system issues: in sequence, dated or random.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: two bookings on property X, one load of PM1's attendee view to see which request it makes, and four further requests; cancel both bookings afterwards. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
I2. Enquiries and bookings: every place they land (read-only)
- Data involved
- An enquirer's contact details and message, and the property they asked about.
- Who should and should not reach it
- Should: the staff who answer enquiries for that property. Should not: every website editor, marketing tools the agency did not choose, or an inbox nobody owns.
- Test case
- From the website's admin settings, the booking tool's settings and any listing site's lead settings, list each destination the settings show for an enquiry or booking: stored entries, notification emails and their recipients, and any customer relationship or property management system it is copied to. Note, too, the website roles permitted to open stored entries.
- Expected result
- The agency chose each destination, a named inbox receives each notification, and stored entries open only for the roles the agency named.
- Evidence to keep
- The destinations and recipients as read, with the date, and the role setting for stored entries.
- Limits of the test
- Read-only: uses the agency's own administrator login on the agency's own accounts only, changes no setting, and opens neither a stored entry nor a connection key or secret. Sends nothing to the form. If a settings screen happens to display a real enquirer's details, stop and note it. What a connected system does with an enquiry after it arrives is outside this row.
R1. Application form: what it asks for (read-only)
- Data involved
- Every field and document request on the rental application form, including any form a listing or application platform adds.
- Who should and should not reach it
- Should: ask only for what the agency's written rule says it needs to assess an application. Should not: collect fields or documents nobody has decided the agency needs.
- Test case
- List every field and document request on each version of the application form the agency uses, and compare the list with the agency's written rule. Where the agency has established with its adviser that Victoria's rules apply to it, also compare the list with the prescribed form as Consumer Affairs Victoria describes it, including its limits on identity documents, financial documents and bank statements with daily transactions.
- Expected result
- Every field and document request maps to a line in the agency's rule, and anything that does not is removed or justified in writing.
- Evidence to keep
- The dated field list for each form version, the rule it was compared with, and the decision for each field that did not match.
- Limits of the test
- Read-only: reads the form and its settings in the agency's own systems and accounts, and submits nothing. If a settings screen happens to display a real applicant's details, stop and note it. Whether a field may be collected under the duties that apply is for the agency and its adviser; the row records the comparison.
R2. Identity and bank documents: who can open them
- Data involved
- Identity documents and bank statements uploaded with an application. In the test, a specimen identity image and an invented bank statement with no transaction details, uploaded by AP1 for property X.
- Who should and should not reach it
- Should: AP1, if the agency lets applicants see their own uploads, PM1, and LL1 only if the agency's rule says so. Should not: other applicants, a landlord of another property, contractors, or anyone signed out.
- Test case
- Note the address or identifier of each specimen file. Request each one as AP2, as AP3, as LL2, as CT1 and with no session (WSTG-ATHZ-02, WSTG-ATHZ-04; ASVS v5.0.0-8.2.2).
- Expected result
- Every one of the ten requests is refused, and a signed-in account gets the same answer it would get for a file that was never uploaded.
- Evidence to keep
- For each request, the account used, the request and the response, plus how the system names each file: a number in sequence, a date or a random string.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: two specimen files, never a real applicant's document, and five requests per file. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
R3. Identity and bank details in lists, shortlists, notifications and exports
- Data involved
- Document numbers, dates of birth, income and account details from an application. In the test, the specimen identity number and the invented account number from R2.
- Who should and should not reach it
- Should: appear only where the agency decided, such as the application record. Should not: application lists, search results, a shortlist sent to the landlord, notification emails or exports, unless the agency decided they should.
- Test case
- As PM1, load the application list for property X and search for AP1, saving each response from the browser's network panel. Send the shortlist to LL1 the way the agency does and read it as LL1. Export the applications for property X only, so the file holds nothing but test records, and read the notification emails AP1's application triggered. Search each of these for the specimen numbers (ASVS v5.0.0-8.2.3).
- Expected result
- The specimen numbers appear only where the agency decided they should, and no response carries a number its screen does not show.
- Evidence to keep
- The saved responses, LL1's copy of the shortlist, the export and the emails, each marked where a specimen number turned up, with any link token redacted.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one request per screen, one shortlist and one export filtered to property X; the export and saved responses go to a folder the agency controls and are deleted once the check is recorded. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
R4. Application reference numbers and status links
- Data involved
- An application's status, the property address and the applicant's name, reached through a reference number or a status link.
- Who should and should not reach it
- Should: the applicant who made the application. Should not: anyone who changes the reference, or anyone signed out, unless the agency decided that the link alone is enough.
- Test case
- Submit applications for property X as AP1 and AP2. Compare their references or status links and record whether they run in sequence or carry a date, address or name. As AP1, request AP2's status by changing the reference, then request AP1's own status with no session. Record the Referrer-Policy header on the status page and whether it loads scripts from third parties (WSTG-ATHZ-04; OWASP Authorization Cheat Sheet).
- Expected result
- AP2's status is refused to AP1. AP1's own status needs a sign-in, or works from the link alone only if the agency decided it should, and then shows the status and nothing more. The Referrer-Policy is one the agency chose.
- Evidence to keep
- Both references with any token redacted, each request and response, and the header value.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: two test applications, one view of each application's own status page and two further requests; tell the leasing staff first so neither application is processed. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
D1. Declined and withdrawn applications: the retention rule and each system's setting (read-only)
- Data involved
- Applications that did not lead to a lease, with their documents, and any copies in email attachments, downloaded files or the platform they came through.
- Who should and should not reach it
- Should: held for the period the agency's written rule sets, and no longer. Should not: still sitting in an application platform, an inbox, a shared drive or a staff member's downloads once that period ends.
- Test case
- Read the agency's written retention rule for unsuccessful and withdrawn applications and for former tenants' information. For each system that holds them, note its retention or deletion setting, then list where copies end up and who is responsible for each. Where the agency has established with its adviser that Victoria's rules apply to it, set the rule beside the timeframes Consumer Affairs Victoria gives.
- Expected result
- Every setting carries out the rule, and someone is named to apply it to each place a copy ends up.
- Evidence to keep
- The written rule, every setting with the date it was read, and the list of copy locations with the person responsible for each.
- Limits of the test
- Read-only: reads a written rule and settings in the agency's own systems, changes nothing and opens no stored application. If a settings screen happens to display a real applicant's file, stop and note it. Whether deletion actually happens is row D2's job, on a test record. The duties that bear on the period are for the agency to settle with its adviser.
D2. After a decline: what the applicant's login and links still reach
- Data involved
- AP2's declined application for property X, its specimen documents and the links AP2 was emailed.
- Who should and should not reach it
- Should: AP2 reaches only what the agency's rule allows after a decline, and nothing once the record is deleted or de-identified. Should not: documents opened through old links, or AP2 found in staff search after deletion.
- Test case
- Decline AP2's application. As AP2, sign in, request each uploaded document by its address and open each link AP2 was emailed. Then run the deletion or de-identification the agency's rule calls for, using the system's own function on this test record, repeat the document requests, and search for AP2 once as PM1 (ASVS v5.0.0-8.2.2).
- Expected result
- After the decline, AP2 reaches only what the rule allows. After deletion, every document address and link returns not found, and the staff search returns nothing for AP2.
- Evidence to keep
- Each request and response before and after deletion with any link token redacted, the time of each step, and the search result.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one test application; one request per document and per link before deletion, the same after, and one search. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
L1. Lease signing links
- Data involved
- The lease, the condition report and any other document sent for signature, reached through a link. In the test, a test lease for property X sent to AP1.
- Who should and should not reach it
- Should: AP1 and the agency, and LL1 if the landlord signs as well. Should not: anyone holding the link after it is used or expires, or anyone who changes the identifier in it.
- Test case
- Send the test lease to AP1. Open the link once and sign. Open it again after signing, once after any expiry the system states, and once with the document identifier changed to another test document. Note the signing page's Referrer-Policy value, and any scripts it pulls in from other sites.
- Expected result
- After signing, the link stops working or shows AP1 only the signed copy, as the agency decided, and it stops after its stated expiry. The changed identifier is refused. The Referrer-Policy is one the agency chose.
- Evidence to keep
- The signing link, token redacted; when it was signed and when it expired; every response; and the Referrer-Policy as served.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one test lease, one signature and three further requests. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
L2. The signed lease, condition report and bond documents: who reaches them
- Data involved
- The signed lease, the condition report with its photos, and the bond paperwork for property X.
- Who should and should not reach it
- Should: AP1 as tenant, PM1 and LL1. Should not: AP2, AP3, LL2, contractors, or anyone signed out.
- Test case
- Note the address or identifier of each document. Request each one as AP2, as AP3, as LL2, as CT1 and with no session. Then, as AP1, request property Y's lease by its identifier (WSTG-ATHZ-02, WSTG-ATHZ-04; ASVS v5.0.0-8.2.2).
- Expected result
- All of them are refused, and to a signed-in account the refusal looks no different from a request for a document that was never created.
- Evidence to keep
- The account behind each request, with the request and its response.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: five requests per document and one for property Y's lease; no edits or deletes. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
M1. Maintenance photo uploads: file type and content
- Data involved
- Photos and files a tenant attaches to a maintenance request.
- Who should and should not reach it
- Should: the maintenance form takes only the file types the agency lists, and keeps them somewhere the server never executes. Should not: a stored attachment that the server executes, or that a browser is handed as active content.
- Test case
- As AP1, the tenant of property X, attach to a maintenance request one file of a type the agency does not accept and one harmless file whose content does not match its extension. If either is stored, fetch it straight from its address and note whether the server executes it and which headers come back with it (WSTG-BUSL-08; ASVS v5.0.0-5.2.2, v5.0.0-5.3.1).
- Expected result
- Both files are rejected, and no stored file is run as server-side code.
- Evidence to keep
- The two upload requests with their responses, the address of anything stored, and the response headers it came back with.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: harmless test files only, never malware samples or a real tenant's photo; two uploads and one direct request per stored file. Remove any stored test file when the row ends, or add it to the clean-up list. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
M2. Maintenance photos: who can open them, and what the served copy carries
- Data involved
- Photos taken inside a home. In the test, one photo taken at the agency's own office, attached by AP1 to a maintenance request for property X.
- Who should and should not reach it
- Should: AP1, PM1, LL1 if the agency shares photos with landlords, and CT1 once assigned the job. Should not: AP3, LL2, CT2, or anyone signed out.
- Test case
- Note the photo's address. Request it as AP3, as LL2, as CT2 and with no session (WSTG-ATHZ-02, WSTG-ATHZ-04; ASVS v5.0.0-8.2.2). Download the copy CT1 receives and the copy LL1 receives, and read each one's metadata for location and device fields beside the original's (OWASP File Upload Cheat Sheet, on removing unnecessary image metadata).
- Expected result
- The four requests are refused. Each served copy keeps only the metadata the agency decided to keep.
- Evidence to keep
- Each request and response, and the metadata of the original and of each served copy side by side.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one test photo taken at the agency's office rather than in anyone's home; four requests and two downloads. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
M3. Maintenance job numbers and tracking links
- Data involved
- A job's status, the property address, the tenant's name and any notes, reached through a job number or a tracking link.
- Who should and should not reach it
- Should: the tenant who reported it, the property manager, the landlord of that property and the assigned contractor. Should not: anyone who changes the job number, or anyone holding the link after the job closes, unless the agency decided otherwise.
- Test case
- Raise one test job for property X as AP1 and one for property Y as AP3. Compare their job numbers or links and record whether they run in sequence. As AP1, request AP3's job by changing the number. After PM1 closes AP1's job, open AP1's link again with no session (WSTG-ATHZ-04; ASVS v5.0.0-8.2.2).
- Expected result
- AP3's job is refused to AP1. AP1's closed job shows no more than the agency decided, or nothing.
- Evidence to keep
- Both job numbers or links with any token redacted, and each request and response.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: two test jobs and two further requests; tell the maintenance team first so no contractor is sent. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
C1. Work orders: each contractor reaches only their own jobs
- Data involved
- Work orders, quotes, photos and invoices for the jobs at properties X and Y.
- Who should and should not reach it
- Should: CT1 for the job at X, CT2 for the job at Y. Should not: either contractor for the other's job, or for any other job at the same property.
- Test case
- As CT2, request the X job by its identifier, its photos and its invoice route, and search or filter the job list for property X. Then, as CT1, do the same for the Y job (WSTG-ATHZ-02, WSTG-ATHZ-04; ASVS v5.0.0-8.2.2).
- Expected result
- Every cross-job request is refused, and no list or search returns the other contractor's job.
- Evidence to keep
- All eight requests with their responses and the contractor account behind each, the refused ones as well.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: two test contractors and two test jobs; four requests in each direction; no edits or deletes. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
C2. Work orders: which tenant details and entry arrangements they carry
- Data involved
- The tenant's name and phone number, the entry arrangements and any other tenancy field on a work order. In the test, an invented placeholder in the entry-arrangements field of the X job.
- Who should and should not reach it
- Should: CT1 receives the fields the agency decided the job needs. Should not: the tenant's email address, the rent, the lease, application details or the landlord's details, unless the agency decided otherwise.
- Test case
- As CT1, open the X work order and save the response behind the screen from the browser's network panel, then open the same work order from the emailed or texted link. Compare the fields each carries with what the screen shows and with the agency's list (ASVS v5.0.0-8.2.3).
- Expected result
- Each response carries only the fields on the agency's list, and no tenant field the screen does not show.
- Evidence to keep
- The saved responses and messages, with any link token redacted and each field outside the agency's list marked.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one work order and two requests. The entry-arrangements field holds an invented placeholder, never a real key safe or alarm code. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
C3. When a job closes or a contractor leaves
- Data involved
- CT1's access to the X job's address, tenant details, entry arrangements and photos.
- Who should and should not reach it
- Should: shrink to what the agency's rule allows when the job closes, and end when CT1 is removed from the agency's contractor list. Should not: survive in a session that was already open.
- Test case
- With CT1 signed in on one browser, close the X job and reload the job in that browser. Then disable or remove CT1's account, reload the job once more in the same browser without signing out, and try to sign in again (ASVS v5.0.0-7.4.2, for the disabled account).
- Expected result
- After closure, CT1 sees only what the rule allows. After removal, the reload and the new sign-in are both refused.
- Evidence to keep
- When the job was closed and when CT1 was removed, each set against the first request refused after it, and every response.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one test contractor and one test job; two reloads and one sign-in. Re-enable CT1 afterwards and record it. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
O1. Owner portals: each landlord reaches only their own property
- Data involved
- Statements, the rent ledger, documents, routine inspection reports and their photos, and maintenance invoices for properties X and Y.
- Who should and should not reach it
- Should: LL1 for property X, LL2 for property Y. Should not: either landlord for the other's property.
- Test case
- As LL2, send again each request LL1's portal makes that names property X or one of its statements, documents, inspection reports or invoices, including list, search and download routes. Then the same as LL1 for property Y (WSTG-ATHZ-02, WSTG-ATHZ-04; ASVS v5.0.0-8.2.2).
- Expected result
- Every cross-owner request is refused with the same response as a record that does not exist, and no list or search returns the other property.
- Evidence to keep
- Every request and response, labelled LL1 or LL2.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: two test landlords on two test properties; one request per route in each direction; no edits or deletes. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
O2. Owner portals: which applicant and tenant fields a landlord receives
- Data involved
- AP1's application, identity and income details, the tenancy record and the routine inspection report for property X, and AP2's declined application.
- Who should and should not reach it
- Should: LL1 receives the fields the agency decided landlords see. Should not: fields outside that list, or anything from AP2's declined application, unless the agency decided otherwise.
- Test case
- As LL1, open the tenancy record, the inspection report and any applicant shortlist for property X, and save the response behind each screen. Compare every field with the agency's list, looking for the specimen numbers from R2 and anything from AP2's application (ASVS v5.0.0-8.2.3).
- Expected result
- Each response carries only the fields on the agency's list, and nothing from AP2 unless the agency decided otherwise.
- Evidence to keep
- The saved responses, with each field outside the list marked.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one test landlord and one request per screen, three at most. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
O3. Owner access when management ends
- Data involved
- Everything LL2 could see for property Y.
- Who should and should not reach it
- Should: end, or shrink to what the agency's rule allows, when the agency stops managing property Y. Should not: continue through a session that was already open.
- Test case
- With LL2 signed in on one browser, end the management of property Y, or remove LL2's portal access, the way the agency does. Reload one statement in that browser without signing out, then try to sign in again (ASVS v5.0.0-7.4.2, where the account is disabled).
- Expected result
- The reload and the new sign-in are refused, or LL2 sees only what the agency's rule allows after management ends.
- Evidence to keep
- When management ended or access was removed, set against the first request refused afterwards, and both responses.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: one test landlord; one reload and one sign-in. Restore the test property and LL2's access afterwards and record it. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
P1. Every place payment details are shown or changed (read-only)
- Data involved
- The agency's own account details on rent notices, tenant pages and templates; each landlord's disbursement account; each contractor's payee details; and any form through which a tenant nominates an account.
- Who should and should not reach it
- Should: only the people and roles the agency named can alter them. Should not: a staff role outside that list, a tenant, or an integration whose job does not involve payment details.
- Test case
- List the templates, pages, portal forms and systems that show or accept payment details. For each, record which roles and integrations can change it, from that system's own role settings, and whether a change takes effect at once or waits for confirmation.
- Expected result
- Someone owns each place, and the roles able to change it are few and match what the agency's rule names.
- Evidence to keep
- The list with the date it was made, and the role setting behind each entry.
- Limits of the test
- Read-only: uses the agency's own administrator login on the agency's own systems only, changes nothing, and leaves bank feed credentials, API keys and payment provider secrets unopened while integration settings are read. If a settings screen happens to display a real landlord's, tenant's or contractor's bank details, stop and note it. How the trust accounting software works inside is outside this row.
P2. A landlord changing their disbursement account
- Data involved
- The account LL1's rent is paid out to. In the test, placeholder account details the accounts team supplies.
- Who should and should not reach it
- Should: LL1 can ask for a change from their own signed-in account. Should not: the change takes effect without the steps the agency's rule requires, or a request names LL2's owner record.
- Test case
- As LL1, change the disbursement account through the owner portal to the placeholder details, then send the same request with LL2's owner identifier in place of LL1's. Check whether LL1 is asked to authenticate again, whether the change is held until staff confirm it, where any notification goes, and which account the staff view shows as in use (WSTG-ATHZ-04, WSTG-IDNT-01; ASVS v5.0.0-2.3.5; OWASP Authentication Cheat Sheet, on re-authentication for critical actions).
- Expected result
- LL1 is asked to authenticate again. The change is held as pending and flagged to staff, a notification goes to the contact details already on record for LL1, and the account in use stays the same until staff confirm through the agency's process and a second staff member approves. The request naming LL2 is refused.
- Evidence to keep
- Both requests and responses, the change as it shows while pending, the notification LL1 received with any link token redacted, the staff screen and the record of the second approval.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: placeholder details only, never a real account; two requests and one approval. Restore the original placeholder and record it. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
P3. A contractor's payee details
- Data involved
- The account CT1's invoices are paid to. In the test, placeholder account details the accounts team supplies.
- Who should and should not reach it
- Should: changed only through the agency's confirmation process. Should not: changed by CT2, or by the account details printed on a new invoice alone.
- Test case
- As CT1, change the payee details through the contractor portal, if the agency has one, then send the same request with CT2's identifier. Separately, upload one test invoice for the X job that shows the placeholder account, and check whether the system changes CT1's payee details or flags the difference to staff (WSTG-ATHZ-04, for the request naming CT2).
- Expected result
- CT1's change is held for confirmation through contact details already on record, CT2's request is refused, and the account on the invoice does not replace the payee details on record without staff confirming it.
- Evidence to keep
- The requests and responses, the invoice as uploaded, the staff view of the payee record and any flag raised.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: placeholder details only; two requests and one invoice upload. Tell the accounts team first so the test invoice is not paid, and restore the payee record afterwards. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
P4. Contact details changed just before payment details
- Data involved
- LL1's email address, then LL1's disbursement account.
- Who should and should not reach it
- Should: confirmation of the payment change reaches contact details that were on record before either change. Should not: the confirmation goes only to the new email address.
- Test case
- As LL1, change the account's email address to a second address the agency controls, then change the disbursement account to the placeholder details. Record whether each change asks LL1 to authenticate again, and which address each notification and confirmation goes to (ASVS v5.0.0-7.5.1, v5.0.0-6.3.7; OWASP Authentication Cheat Sheet, on changing a registered email address).
- Expected result
- The email change asks LL1 to authenticate again and sends a message to the original address. The payment change is held, and its confirmation goes to the contact details on record before the email change.
- Evidence to keep
- Each request and response, and each message as received at each address with any link token redacted, with times.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: placeholder details and two addresses the agency controls; two changes. Restore both afterwards and record it. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
P5. The agency's own account details on rent notices and tenant pages
- Data involved
- The account name, BSB and account number tenants pay rent into, as shown on templates and tenant payment pages.
- Who should and should not reach it
- Should: one staff role the agency names puts the change forward, and a different named person signs it off. Should not: a staff member outside that role, a tenant, or any single person changing the details unaided.
- Test case
- On a staging copy only, try the change as a property manager account in a role the agency has not named for payment details, and then as an account in the role it has named, first through the screen and then by sending the request directly, with placeholder details the accounts team supplies. For the permitted attempt, note whether the new details go live before anyone else signs them off, whether a fresh sign-in or other re-authentication is asked for, and which people are notified (WSTG-IDNT-01; ASVS v5.0.0-2.3.5).
- Expected result
- The unnamed role is refused. The permitted change takes effect only after every step the agency's rule requires, and it is logged with who made it and when, and the people the agency named are notified.
- Evidence to keep
- All four attempts with their responses, the sign-off record, the log line for the change and the notifications as they arrived.
- Limits of the test
- Sends requests: only systems the agency owns or controls, on a staging copy where one exists, with the invented test roles and records; no load; any email or text goes to an address or number the agency controls; a link that opens a record without a sign-in is opened only for the invented records, and its token is redacted from every piece of evidence; session cookies and tokens are removed from every request and response kept; if a vendor runs the system, read the vendor's terms and ask the vendor first (row V1), and run the row there only as far as those terms and the vendor's written reply allow. Requests in this row: a staging copy only, never live payment details; two attempts per role. Put the original placeholder values back and note when that was done. If a response shows a real person's data, stop, record where and when, open nothing further and tell whoever handles privacy for the agency.
V1. Third-party systems in the rental (read-only)
- Data involved
- What the agency's property management software, application platform, booking tool, e-signing service, owner portal and payment services hold, and what each integration between them can reach.
- Who should and should not reach it
- Should: an integration can reach the tenancies, owners or jobs its purpose calls for and nothing beyond them. Should not: a connection set up for one task, such as sending inspection reminders, that can read every tenancy and owner record.
- Test case
- Before any row above touches a vendor's system, find what that vendor's terms say about security testing and put the question to the vendor in writing. Then, in the agency's own admin settings, record every connection between these systems, which account each one runs as and what each can reach.
- Expected result
- No connection reaches beyond its purpose, and a written reply from each vendor is on file before any row runs inside that vendor's system.
- Evidence to keep
- Each vendor's terms as read, each written reply, and the connection list with the date it was made.
- Limits of the test
- Read-only: reads terms and the agency's own settings and writes to the vendor; changes nothing, and leaves every key, token and secret unopened. If a settings screen happens to display a real tenant's or landlord's details, stop and note it. The checklist says nothing about what a vendor allows, and testing a vendor's own platform is for that vendor.
Real estate cyber security, followed through a rental
In a property management business, personal information can reach the web at each step of a rental.
Someone books an inspection through a listing or the agency's website. They apply, and an application can carry identity documents and evidence of income. The chosen applicant signs a lease and any condition report, the others are told no, and the question of what happens to their documents begins.
Through the tenancy, the tenant may report repairs with photos taken inside their home, contractors may receive work orders that say how to get in, and the landlord may follow it all through statements and reports.
Money moves the whole time: rent comes in, disbursements go out to the landlord, and contractors' invoices are paid.
Each step involves a different person who should see a different part of that data: the applicant, the property manager, the landlord and the contractor. Real estate agency cyber security, on the web side, is the question of whether the agency's systems keep each of those people to their own part.
Property management website security is the same question asked of the forms, portals and links the agency runs. This page follows the steps in order, names the data at each one and who should and should not reach it, and the rental-to-lease checklist turns each step into tests to run on the agency's own systems, or to hand to whoever looks after them.
It is built around tenancy documents, maintenance access and changes to payment details, because at those points the data is a person's identity, the inside of their home or the destination of money.
It is written for the principal, licensee, property management lead or technical lead of an Australian agency. It is not legal advice. It quotes the laws and regulator pages it relies on, says who they cover, and leaves the question of whether each applies to the agency with the agency and its own adviser.
Where tenant, landlord and contractor data moves in a rental
The subsections follow the checklist's row groups in order.
Each says what the step involves and where access can go wrong, and names the guide in this library that explains the weakness, so this page can stay with the agency's own workflow.
An inspection booking can be the first time a prospective renter gives the agency a name, a phone number and an email address, tied to a property they want. If the booking tool sends a confirmation with a link or reference for managing the booking, that reference is the thing to test: whether changing it shows another person's booking, and whether the list of everyone booked into a time slot can be fetched by anyone other than the property manager.
Both are object-level access checks, which the broken access control and IDOR guide explains in depth.
Row I2 has the agency list the places an enquiry lands, such as stored website entries, notification emails and any listing site or customer relationship tool it is copied to, since each copy has its own readers.
Rental application data security turns on three questions. Who can open an uploaded identity document or bank statement? Where else do the numbers on them appear: application lists, search results, a shortlist sent to the landlord, notification emails, exports? And does the form ask for more than the agency has decided it needs?
On the third, and for Victoria only, Consumer Affairs Victoria says rental providers and their agents must use the prescribed form for rental applications, and that a rental provider or agent can only ask for the information set out in it, limited to what is necessary to decide if a renter is suitable, for example to confirm their identity and capacity to pay the advertised rent.
It says a rental provider or agent may request no more than two documents from pay slips from current or recent employment, a bank statement without transaction details and Centrelink payment statements or letters to confirm capacity to pay the rent, and no more than two identity documents from a list it gives, and that the rental provider may ask an applicant who is not receiving regular income for details of previous employment or other documents supporting their financial ability to pay rent, such as proof of savings or assets. The section on laws and regulator guidance says how far this page goes with tenancy law.
The first two questions are access checks. In ASVS 5.0, requirement 8.2.2 (Level 1) asks that data-specific access is restricted to consumers with explicit permissions to specific data items, and requirement 8.2.3 (Level 2) asks that field-level access is restricted to consumers with explicit permissions to specific fields.
How a list or an export comes to carry fields its screen does not show is explained in the sensitive data exposure guide. How the upload itself is handled, with the extension and the content both checked and the file stored where the server will not run it, belongs to the file upload security guide, and row M1 applies its tests to maintenance photos.
If the agency's application tool gives each applicant a reference number, or a link that shows progress without a sign-in, that identifier works like a key to a record: whoever holds it, or works out the next one, reaches the status page.
OWASP's Authorization Cheat Sheet says masking or randomising lookup identifiers is generally not sufficient by itself. So row R4 asks two separate questions: whether the server checks that the person asking owns the application, and what the status page shows to someone who reaches it anyway.
A reference that counts up by one, or carries the property address or the applicant's name, makes the second question matter more. Maintenance job numbers and tracking links get the same pair of tests in row M3.
When a lease is signed, any other applications for the property are left behind, each with its documents. What happens to them is a retention decision with two halves: the agency's rule for how long it keeps them, and whether its systems carry that rule out, including the copies in email attachments, downloaded files and any platform the application came through.
For Victoria, Consumer Affairs Victoria gives timeframes for destroying or permanently de-identifying a renter's information, including timeframes for applicants who were unsuccessful, and APP 11.2 of the Privacy Act 1988 deals with personal information an APP entity no longer needs. Both are quoted in the section on laws and regulator guidance below.
The web-side test is narrower than either. Once an application is declined or has expired, does the applicant's login still open their documents? Do the links they were emailed still work? And after the agency's deletion or de-identification step runs on a test record, does any document address, link or staff search still return it? Rows D1 and D2 check the rule and then the systems.
The lease, the condition report and any bond paperwork are documents for the tenant, the landlord and the agency. Where they go out as links in an email, the link decides who can open them. A link that opens a document without a sign-in is open to whoever holds it, so the tests ask when each link stops working: after it has been used to sign, after any expiry the system states, and when the identifier in it is changed.
Where a session ID travels in the URL, OWASP's Session Management Cheat Sheet says the mechanism might disclose it in web links and logs, web browser history and bookmarks, the Referer header or search engines.
A link that opens a document without a sign-in carries its token in the URL too, which is why rows R4 and L1 record the Referrer-Policy header and any third-party scripts on the page a link opens.
If signing runs on an e-signing service, the vendor's platform is the vendor's, and the agency's tests stay with its own settings and its own side of the integration, under the vendor's terms.
A maintenance request can include photos taken inside someone's home: a leak under the sink, mould in a bedroom, a broken lock on the back door. Those photos can show belongings, papers left on a bench or the people who live there, and they can pass from the tenant to the property manager, the landlord and the contractor who quotes the job.
Each copy is a file whose download route needs its own access check, because an uploaded file can be served from a different route from the screen that shows it.
The file upload security guide covers what an upload point should accept. The checklist applies that to maintenance photos and adds two questions of its own: whether another tenant, another landlord or an unassigned contractor can open a photo by its address, and whether the copy the system serves keeps any metadata the original carried, which row M2 checks for location and device fields.
For images, OWASP's File Upload Cheat Sheet advises decoding and re-encoding to an allowed image format, explicitly removing unnecessary metadata, and says that rewriting does not guarantee that all malicious content is removed.
A work order is the contractor's view of a tenancy. To do the job, a contractor may need the address, the tenant's name and phone number, and how to get in. Unless the agency decides otherwise, the job does not call for the tenant's email address, the rent, the lease, the application, or jobs assigned to other contractors.
OWASP's Authorization Cheat Sheet says that just because a user has access to an object of a particular type does not mean they should have access to every object of that type. A contractor is in that position: access to jobs, but not to every job.
So the questions are horizontal (can one contractor open another contractor's job by changing its number?), field-level (does the work order response carry tenant fields the screen does not show?) and then about time. When a job closes, or a contractor is taken off the agency's list, does their access to the address, the tenant's details and the entry arrangements end, including through a session that was already open?
Where taking a contractor off the list disables their account, ASVS 5.0 asks at Level 1 that the application terminates all active sessions when a user account is disabled or deleted (requirement 7.4.2). Session expiry and sign-out in general are left to the session management guide.
Real estate client portal security, from the landlord's side, starts with two owners. Can one landlord's login reach another landlord's statements, rent ledger, documents or inspection reports by changing an identifier, in either direction? WSTG v4.2's insecure direct object reference test says the best way to test for direct object references would be by having at least two (often more) users to cover different owned objects and functions (WSTG-ATHZ-04). For the owner portal, the checklist's users are the two test landlords, each owning one test property.
The second question is what tenant data a landlord's view carries. A routine inspection report can show the inside of a tenant's home, and a shortlist can show an applicant's details. Which of those a landlord sees is the agency's decision, and row O2 compares what the owner portal returns, field by field, with that decision.
The third question is when access ends, for example when the agency stops managing the property. If the owner portal is part of a vendor's property management software, the agency reviews its own settings and test accounts inside it, under the vendor's terms.
The checklist's change tests cover three kinds of payment detail: the agency's own account details on rent notices and tenant pages, the account each landlord's disbursements are paid into, and the account each contractor's invoices are paid to, and row P1 also lists any form through which a tenant nominates an account.
A change to any of them, once acted on, changes where money goes. So rows P2 to P5 ask, between them, who can make the change, whether it waits for confirmation or a second person's sign-off, whether the person is asked to authenticate again, who is told and whether the change is logged, and each row says which of these it checks.
Among the times to trigger re-authentication, OWASP's Authentication Cheat Sheet names critical actions: high-risk actions like changing payment details or adding new trusted devices. ASVS 5.0 requirement 2.3.5, at Level 3, asks that high-value business logic flows require multi-user approval to prevent unauthorised or accidental actions, and says this could include but is not limited to large monetary transfers.
On administrators, WSTG v4.2's role definitions test (WSTG-IDNT-01) says sensitive admin functionality should leverage a maker-checker principle, or use MFA to ensure that the administrator is conducting the transaction.
Contact details matter as much as the account number. If confirmations go to whatever details are on the account at the time, a change of email address or phone number made just before a change of bank account sends the confirmation to whoever made the first change.
ASVS 5.0 requirement 7.5.1, at Level 2, is that the application requires full re-authentication before allowing modifications to sensitive account attributes which may affect authentication, such as email address, phone number, MFA configuration or other information used in account recovery.
The Authentication Cheat Sheet's recommended process for changing a registered email address sends a message to the current address as well as the proposed new one, whether or not the user has MFA enabled.
Row P4 tests that sequence. The business logic vulnerabilities guide covers turning a rule such as "nothing is paid to a changed account until it is confirmed" into a test case.
Four test roles, and the agency rules they check
A row can only pass or fail against a rule, so the agency writes down who may reach what before the first request goes out.
At Level 1, ASVS 5.0 asks that authorisation documentation defines rules for restricting function-level and data-specific access based on consumer permissions and resource attributes (requirement 8.1.1). For an agency, those rules can be short, one paragraph for each role:
- Applicant: their own application, its documents and its status, for as long as the agency's retention rule allows, and nothing belonging to any other applicant. After a lease is signed, the approved applicant becomes the tenant.
- Property manager: the properties in their portfolio, if the agency limits managers to their own portfolio, with every change to payment details going through the agency's approval rule.
- Landlord: their own properties' statements, ledger, documents and reports, and only the tenant and applicant fields the agency decides landlords see.
- Contractor: the jobs assigned to them, the fields those jobs need, and only while each job is open.
Invented people for each role
The objectives WSTG v4.2 sets for its role definitions test (WSTG-IDNT-01) are to identify and document roles used by the application, attempt to switch, change or access another role, and review the granularity of the roles and the needs behind the permissions given. For the checklist that means invented people in every role: two applicants for test property X, one approved, who becomes its tenant, and one declined; a third applicant who is the tenant of test property Y; two property managers, one for each property; two landlords, each owning one of the properties; and two contractors, each assigned a job at one of them.
Each has an email address and phone number the agency controls, and every identity document and bank statement they upload is made for the test and marked as a specimen.
Where an answer is not written down, settle it before testing. Whether a landlord sees the declined applicants for their property, whether a contractor sees the tenant's phone number or only an appointment time, and how long a declined application is kept are decisions for the agency, made with the duties below in mind.
Each decision becomes a row's expected result, and the row's evidence shows whether the agency's systems behaved that way.
Laws, regulator guidance and outside requests an agency may weigh
This section repeats what each law or regulator page says, and the people or businesses it names, without drawing a conclusion for any agency.
An agency settles whether a provision reaches it, and what it would ask of that agency, with its own adviser, the regulator concerned or its professional body.
The Australian Privacy Principles are set out in Schedule 1 of the Privacy Act 1988. In section 6(1), APP entity means an agency or organisation. The definition of organisation in section 6C(1) leaves out a small business operator, along with a registered political party, an agency, a State or Territory authority and a prescribed instrumentality of a State or Territory.
Under section 6D(1), a business is a small business at a time in a financial year if its annual turnover for the previous financial year is $3,000,000 or less. Section 6D(2) gives a separate test for a business that was not carried on at any time in the previous financial year.
Section 6D(3) says a small business operator is an individual, body corporate, partnership, unincorporated association or trust that carries on one or more small businesses and does not carry on a business that is not a small business, and section 6D(4) lists cases in which an entity is not a small business operator despite that.
Subsections 6D(5) to 6D(8) qualify some of those cases, and section 6D(9) says a body corporate is not a small business operator if it is related to a body corporate that carries on a business that is not a small business.
Further cases in which a small business operator is treated as an organisation are in sections 6E and 6EA. Among them, section 6E(1A) says that if a small business operator is a reporting entity or an authorised agent of a reporting entity because of anything done in the course of a small business it carries on, the Act applies, with the prescribed modifications (if any), in relation to its activities for the purposes of, or in connection with, activities relating to the Anti-Money Laundering and Counter-Terrorism Financing Act 2006 or the regulations or AML/CTF Rules under it, as if it were an organisation. Working through those provisions for a particular agency is a job for its adviser.
For an APP entity that is an organisation, APP 3.2 says the entity must not collect personal information (other than sensitive information) unless the information is reasonably necessary for one or more of the entity's functions or activities. Row R1 lists what each application form asks for, which gives an adviser the list to weigh against that principle.
APP 11.1 says that if an APP entity holds personal information, it must take such steps as are reasonable in the circumstances to protect the information from misuse, interference and loss, and from unauthorised access, modification or disclosure, and APP 11.3 says those steps include technical and organisational measures.
APP 11.2 says that where an APP entity holds personal information about an individual, no longer needs it for any purpose for which it may be used or disclosed by the entity under Schedule 1, the information is not contained in a Commonwealth record, and the entity is not required by or under an Australian law, or a court/tribunal order, to retain it, the entity must take such steps as are reasonable in the circumstances to destroy the information or to ensure that it is de-identified. The APP 11 guide in this library goes further into that principle.
The OAIC's page on the Notifiable Data Breaches scheme says an organisation or agency covered by the Privacy Act 1988 must notify affected individuals and the OAIC when a data breach is likely to result in serious harm to a person whose personal information is involved.
Should a row turn up a real applicant's documents, a real tenant's photos or real bank details open to someone outside the agency's rule, the tester records it and the agency's privacy adviser decides whether the scheme is engaged.
The map of rental data the checklist needs can also feed a privacy impact assessment when the agency plans a new application platform or owner portal; both the scheme and the assessment have guides in this library.
The only state regulator quoted on this page is Consumer Affairs Victoria, and everything below is credited to it and speaks for Victoria. The page makes no statement about the tenancy laws of other states and territories. Which tenancy laws apply to the agency is for it to check with its adviser.
Consumer Affairs Victoria says renter's information is any information disclosed by a renter or a prospective renter to the rental provider or agent when applying for or entering into a rental agreement or for the purposes of administering the rental agreement. It says rental providers and agents must take reasonable steps to protect renter's information from misuse or loss, and from unauthorised access, modification or disclosure, and that rental providers and agents who hold a renter's information must destroy or permanently de-identify it within specific timeframes.
It introduces its table as the scenarios and timeframes the rental provider or agent must wait before they destroy or de-identify a renter's information, and each timeframe in the table sits beside its scenario: within 3 years after the end date of the rental agreement, where the renter's agreement has ended; within 30 days after the rental agreement is entered into, where an applicant was unsuccessful; and within 6 months after the initial rental agreement they applied for is entered into, where an unsuccessful applicant has provided written consent for the information to be used to apply for other premises.
It describes some limited situations in which the information need not be destroyed or permanently de-identified, or may be disclosed, among them where the rental provider or agent has obtained written consent from the renter to disclose the information and holds it for that purpose, and where an Act, law or order of a court or Tribunal otherwise requires the rental provider or agent to hold or disclose it.
It also says a rental provider or agent is not required to comply with these obligations if they are complying with a corresponding obligation under the acts it lists, which include the Privacy Act 1988 of the Commonwealth. For the law itself, it points to sections 505BA to 505BE of the Residential Tenancies Act 1997.
On applications, its page on applying for a property says a rental provider or agent can only ask for the information set out in the prescribed rental application form, and that rental providers and agents can only use a renter's personal information to decide whether or not they are a suitable renter, or to comply with another requirement of the Residential Tenancies Act 1997.
Among the things it says cannot be asked for is a bank statement with daily transactions, noting that the applicant may need to provide a statement but can delete transactions they feel should be kept private.
It also says that a rental provider or agent who uses tenancy databases to check rental history needs to tell the applicant the names of the databases they will use.
Rows R1 and D1 are the ones to read beside these pages, where the agency has established that they apply to it.
The compilation of the Act quoted here shows the law as amended and in force on 1 October 2026. For the purposes of the Act, the tables in its section 6 define the provision of a designated service and the person (the customer) to whom the designated service is provided.
Table 5 is headed Real estate services. Its item 1 is brokering the sale, purchase or transfer of real estate on behalf of a buyer, seller, transferee or transferor in the course of carrying on a business, and its item 2 is selling or transferring real estate in the course of carrying on a business selling real estate, where the sale or transfer is not brokered by an independent real estate agent.
Table 6 is headed Professional services. Its item 1 is assisting a person in the planning or execution of a transaction, or otherwise acting for or on behalf of a person in a transaction, to sell, buy or otherwise transfer real estate, where the service is provided in the course of carrying on a business, and the sale, purchase or other transfer is not pursuant to, or resulting from, an order of a court or tribunal.
Table 6 has further items, and so do the other tables in section 6, which the agency's adviser can read against the services the agency provides.
Section 6(6) then sets a geographical condition. It says an item of a table in section 6, other than item 32A of table 1, does not apply to the provision by a person of a service to a customer unless a geographical link it sets out is present, such as the service being provided at or through a permanent establishment of the person in Australia.
Section 6 also has subsections, such as 6(6A) on services within business groups, that say when a service is not a designated service despite anything in the section.
This page does not quote AUSTRAC's guidance, and makes no statement about which of an agency's services, if any, is a designated service, or from when. That is for the agency to check against AUSTRAC's guidance with its adviser. Where the answer is yes for any service, the identity-document rows (R2 and R3) and the payment-detail rows (P1 to P5) are the ones to read alongside that advice.
A landlord with a large portfolio, a franchise group or a cyber insurer may send the agency a security questionnaire. Where a question touches applications, owner portals or payment changes, the checklist's dated evidence is something to answer from rather than memory.
For the answers themselves, the security questionnaires guide shows how to tie each one to its evidence, and the cyber insurance security checklist deals with a proposal form's web application questions.
How to run the rental-to-lease checklist
The checklist has twenty-five rows in nine groups.
Eight follow a rental from the first inspection booking to the payments made during the tenancy, and the ninth covers third-party systems. Across its seven columns, a row names the step and check, the data, the people who should and should not reach it, the test case, the result the agency expects, the evidence to keep and the row's own limits.
- Keep to systems the agency owns or controls, and use a staging copy if the agency has one. Before a row touches a system a vendor runs, such as property management software, an application platform, a booking tool, an e-signing service or a payment service, read that vendor's terms and ask the vendor. What a vendor permits is for its own terms and its own answer to say; this page does not say it.
- Set up the invented test roles described above, with test properties X and Y and contact details the agency controls. Make the identity documents and bank statements for the test and mark each as a specimen. Do not upload a real applicant's file, a past one included.
- Fill in each row's expected result from the agency's written rules first. The checklist's expected results are examples, and the agency's own rule replaces any example it disagrees with.
- Send only the requests each row names. Nothing here is a load test, and no row guesses at identifiers in bulk: rows R4 and M3 compare two references the test itself created.
- Tell the property managers and the accounts team before you start, so that a test booking, application, job or change of bank account is not acted on as if it were real.
- A real applicant, tenant, landlord or contractor appearing in any response ends the row. Write down what came back, the URL and the time, request nothing more, and hand the note to the agency's privacy contact.
- Five rows only read and submit nothing: I2, R1, D1 and P1 read a form, settings, a rule or a list, and V1 reads a vendor's terms and puts a question to the vendor. None of the five changes anything. No row asks anyone to open a screen or file that shows a password, key or secret, and the entry-arrangements field in row C2 holds an invented placeholder rather than a real code.
- Treat any link that opens a record without a sign-in, such as a booking, status, signing or tracking link, as a secret. Open it only for the invented records, and redact its token from every request, response, email and screenshot kept as evidence.
- Undo what each row changes: cancel test bookings, delete test documents and photos, restore placeholder payment details, re-enable test accounts, and list anything that could not be removed.
- File the evidence from rows that pass beside the evidence from rows that fail. A dated refusal shows that a check ran; without it, a pass is only a memory.
- Run the checklist again after a change of property management software, application platform or owner portal, after a new integration, and when the agency starts collecting a new kind of document.
Where the checklist stops
A row that passes is a result for the test accounts, test properties and build it ran against, on the day it ran.
It says nothing about routes nobody listed or about the next release. Some of an agency's exposure sits outside any web test altogether.
- Vendor platforms. Whether a vendor's platform enforces its own rules for every agency on it is the vendor's to test. A row run inside a vendor's system speaks only for this agency's instance and the test roles it set up there.
- Email, texts and calls. Payment instructions and documents also travel by email, text message and phone. A message from a compromised mailbox, or a call from someone claiming to be a landlord, does not pass through the portal. Email accounts, staff phones and the office network sit outside both a website audit and a web application test.
- Keys and entry. Keys, key safes and alarm codes are physical access to a home. The checklist tests whether a contractor's login still shows the entry arrangements after a job closes, and it cannot show who holds a key.
- Authorised people. Someone with a valid login who copies data out is acting within the access the rules gave them. The checklist tests the rules, not what an authorised person does with that access.
- Process steps. A call to a number already on record before a changed bank account is used is a process. The checklist can record that the step exists, but it cannot test that people follow it.
- Legal questions. The checklist cannot say whether a provision quoted here reaches the agency, or whether the agency's practice satisfies it: those are questions for its adviser. Its evidence shows how the systems responded, with dates.
Where a Clever Ops test fits for an agency
Clever Ops tests what customers reach through a browser or an API.
Which of its services suits an agency turns on what the agency runs itself rather than licenses. An agency website made of listings and enquiry forms fits the website security check, an external review of what anyone on the internet can see without logging in: certificates and transport security, security headers, cookie settings, exposed files and admin paths, and software versions visible from outside.
Where that website is built on WordPress, the WordPress security audit goes further, following each role that can log in and checking the login and password-reset flows and the installed plugin and theme versions, with evidence.
An application of the agency's own that applicants, tenants, landlords or contractors sign in to calls for the web application and API penetration test, a signed-in test that covers authentication and sessions, what each role can see and change, whether one customer can reach another customer's records, the API called directly rather than only through the screens, payment and billing flows, file uploads and the business rules the application is meant to enforce.
An AI chatbot on the agency's website that answers rental enquiries is tested under the AI chatbot and agent security test, which tests the chatbot, not the AI model behind it.
For that penetration test, the scope asks for two test accounts in each role, so that two landlords or two contractors can each try to reach the other's records, and a staging copy that matches production is preferred.
Destructive actions are off by default for real people's accounts and records that existed before the test, and reset, lockout and deletion checks run only on test accounts and test records where the rules of engagement allow them.
Evidence has session tokens, passwords, keys and personal details redacted, is stored encrypted in Australia, and is deleted on the schedule set in the data handling terms.
Clever Ops does not test a vendor's own product or platform. Property management software, an application platform, an owner portal or an e-signing service that the agency licenses stays with its vendor. A scope can name the agency's own website or application, including that application's side of each integration with a vendor's system.
A test report is evidence for a decision someone else makes, such as a landlord, an insurer or a franchise group, and it is not a certificate or a compliance opinion. Clever Ops holds no security certification or accreditation. Where a contract, panel or tender requires a CREST-accredited provider, the work goes elsewhere: Clever Ops is not CREST accredited, so we refer that work to a provider that is.
Where Clever Ops built or hosts the agency's website or application, the same work is a security review rather than an independent test, and the report names who built it and who hosts it.
Testing of the office network, staff devices or phishing belongs with a network or red team specialist. An agency in the middle of an incident should call an incident response specialist before booking any test.
Frequently asked questions
It turns on facts this page cannot see, and it is a question for the agency's adviser. In section 6C(1) of the Privacy Act 1988, the definition of organisation excludes a small business operator. Under section 6D(1), a business is a small business at a time in a financial year if its annual turnover for the previous financial year is $3,000,000 or less, with a separate test for a new business in section 6D(2), the definition of a small business operator in section 6D(3), and the cases in section 6D(4), as qualified by sections 6D(5) to 6D(8), and in section 6D(9) in which an entity is not a small business operator. Sections 6E and 6EA set out cases in which a small business operator is treated as an organisation, among them section 6E(1A), which says that if a small business operator is a reporting entity or an authorised agent of a reporting entity because of anything done in the course of a small business it carries on, the Act applies, with the prescribed modifications (if any), in relation to its activities for the purposes of, or in connection with, activities relating to the AML/CTF Act or the regulations or AML/CTF Rules under it, as if it were an organisation. Victoria's rules on renters' information, as Consumer Affairs Victoria describes them, are a separate question.
As long as the agency's own written rule says, and that rule is for the agency and its adviser to set against the duties that apply to it. For Victoria, Consumer Affairs Victoria says rental providers and agents who hold a renter's information must destroy or permanently de-identify it within specific timeframes, and for an unsuccessful applicant it gives within 30 days after the rental agreement is entered into, or, where the applicant has provided written consent for the information to be used to apply for other premises, within 6 months after the initial rental agreement they applied for is entered into. It also describes limited situations in which the information need not be destroyed or permanently de-identified. For an APP entity, APP 11.2 of the Privacy Act 1988 deals with personal information the entity no longer needs, on the conditions it sets out. Rows D1 and D2 check that the agency's systems carry its rule out, including the document links a declined applicant still holds.
That is the agency's decision, made with its adviser, and the checklist tests that the owner portal keeps to it. For Victoria, Consumer Affairs Victoria says rental providers and agents can only use a renter's personal information to decide whether or not they are a suitable renter, or to comply with another requirement of the Residential Tenancies Act 1997. Whatever the agency decides, write it down field by field: which applicant and tenant details a landlord sees on a shortlist, in the tenancy record and in inspection reports. Row O2 compares what the owner portal returns, including fields its screens do not show, with that list.
The agency sets the rule, and the checklist tests that its systems hold to it. For a landlord's disbursement account, the rows assume four parts: the change comes from the landlord's own signed-in account; the landlord is asked to authenticate again; the change is held until staff confirm it through the agency's process, with its confirmation going to contact details that were on record before the change; and a second staff member approves it before the account in use changes. The last part matches ASVS 5.0 requirement 2.3.5 (Level 3), which asks that high-value business logic flows require multi-user approval to prevent unauthorised or accidental actions, and says this could include but is not limited to large monetary transfers. Rows P2 and P4 test those parts between them. For a contractor's payee details, row P3 assumes the change is held for confirmation through contact details already on record, and that an account printed on a new invoice does not replace the payee details on record unless staff confirm it. The confirming call itself is a process, which the checklist can record but not test.
If the vendor's system offers contractor logins, how they work is for the vendor to describe, and Clever Ops does not test a vendor's own product or platform. What the agency can check in its own instance, once it has read the vendor's terms and asked the vendor (row V1), and only as far as the terms and the vendor's answer allow, is that each contractor reaches only the jobs assigned to them, sees only the fields those jobs need, and loses access to the address, the tenant's details and the entry arrangements when a job closes or they leave the agency's list. Rows C1 to C3 test those three things with two test contractors, and the entry-arrangements field holds an invented placeholder throughout.
Scope the right test for your application
Tell us what you have built, who uses it and when it needs to launch. You will get a plain-English recommendation: a website audit, a web application test or a referral to a specialist.