IDOR in a Banking App: What a Good Bug Bounty Report Looks Like
A good IDOR report for a banking application states, in the first lines, which endpoint lets one customer access another customer's data or actions, shows the exact requests using two test accounts the researcher owns, explains the impact in banking terms (statements, balances, payees, transfers), and gives a clear fix: check that the account or object belongs to the authenticated customer on the server.
Topic: Security Tools, Authorized Labs & CTFs in Cybersecurity & Ethical Hacking · Explainer
Also covers: Web & API Security
Quick answer: A good IDOR report for a banking application states, in the first lines, which endpoint lets one customer access another customer's data or actions, shows the exact requests using two test accounts the researcher owns, explains the impact in banking terms (statements, balances, payees, transfers), and gives a clear fix: check that the account or object belongs to the authenticated customer on the server. It contains only the minimum data needed to prove the issue. The example below is illustrative; it is not a real bank's vulnerability.
Key takeaways
- IDOR in banking usually appears in statement downloads, account summaries, payees and transaction APIs.
- Prove it with two accounts you control, never with real customers' data.
- A strong report has a one-line summary, steps, evidence, impact, severity and fix.
- Write-actions (adding payees, moving money) are far more severe than read-only leaks.
- Only test banks that run a published bug bounty or vulnerability disclosure programme.
Table of Contents
- What Is IDOR in a Banking Context?
- Where Does IDOR Appear in Banking Apps?
- What Does a Sample Report Look Like?
- How Do You Rate Severity?
- How Should the Bank Fix It?
- Why Do IDOR Reports Get Rejected?
- How Do You Practise and Report Legally in India?
- References
What Is IDOR in a Banking Context?
An insecure direct object reference happens when the server uses an identifier from the request, such as an account number or statement ID, without confirming it belongs to the logged-in customer. OWASP lists this as API1: Broken Object Level Authorization, and MITRE as CWE-639. See our broken access control guide for the wider family.
Where Does IDOR Appear in Banking Apps?
| Feature | Typical parameter | Impact if unchecked |
|---|---|---|
| Statement download | statementId, accountNo | Full transaction history exposed |
| Account summary API | customerId | Balances, KYC details |
| Payee management | payeeId | Add or delete payees on another account |
| Fund transfer | fromAccount | Debit another customer's account |
| Card controls | cardId | Block, unblock or change limits |
| Support tickets and uploads | ticketId, file path | KYC document leakage |
Mobile banking apps often call APIs directly, so testers inspect the app's traffic through a proxy on a test device.
What Does a Sample Report Look Like?
Illustrative example. "ExampleBank" and all values are fictional.
Title: IDOR in /api/v2/statements/{id}/pdf lets any customer download
another customer's account statement
Summary
The statement download endpoint returns a PDF for any statement ID without
checking it belongs to the authenticated customer.
Asset / scope: api.examplebank.test (in scope, "Mobile banking API")
Steps to reproduce (two researcher-owned test accounts)
1. Log in as Customer A (acct ending 1111). Open Statements, download June.
Request: GET /api/v2/statements/880214/pdf -> 200, A's statement
2. Log in as Customer B (acct ending 2222) and note B's statement ID 880215.
3. As Customer A, request:
GET /api/v2/statements/880215/pdf
Authorization: Bearer <A's token>
-> 200 OK, PDF of Customer B's June statement
4. Statement IDs are sequential, so other customers' IDs are predictable.
I did not request any ID other than my own two accounts.
Impact
Any logged-in customer can download other customers' statements: names,
account numbers, balances and full transaction history.
Severity: High. CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
Suggested fix
Look up the statement by id AND the authenticated customer's account list;
return 404 if it does not belong to them. Consider non-sequential IDs as
defence in depth, and add an automated test for cross-account access.
Evidence: two screenshots and the raw HTTP requests (attached, redacted).
How Do You Rate Severity?
Use the programme's scale, often CVSS. Points that raise severity in banking:
- Write access (transfers, payees, card controls) is usually critical.
- Predictable IDs make mass extraction possible.
- Sensitive data such as KYC documents or full transaction history.
- No user interaction needed from the victim.
State facts and let the triage team decide; overstated severity reduces trust.
How Should the Bank Fix It?
// Express.js example: scope the lookup to the customer's own accounts
app.get('/api/v2/statements/:id/pdf', requireAuth, async (req, res) => {
const stmt = await db.statement.findFirst({
where: { id: Number(req.params.id),
account: { customerId: req.user.customerId } }
});
if (!stmt) return res.sendStatus(404);
res.type('application/pdf').send(await renderPdf(stmt));
});
Banks should also log and alert on one customer requesting many different object IDs, and include authorization tests in release pipelines. In India, the RBI's IT governance direction expects regular security testing of internet and mobile banking.
Why Do IDOR Reports Get Rejected?
- The "other" object belongs to the same user or is public by design.
- The report uses real customers' data, breaching programme rules.
- No proof the response actually contained another user's data.
- Out-of-scope domain or app version.
- Vague impact, such as "could lead to data breach", with no evidence.
HackerOne's quality report guidance is a useful checklist.
How Do You Practise and Report Legally in India?
Practise on deliberately vulnerable apps such as OWASP Juice Shop or crAPI. Test real banks only through their published bug bounty or vulnerability disclosure programme. If a bank has none, report through CERT-In rather than testing further; unauthorised access can breach Sections 43 and 66 of the IT Act 2000. Read a beginner's first bug journey, or ask WebAsha about bug bounty training.
References
Frequently Asked Questions
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Wow
0
Sad
0
Angry
0