CVE-2026-8785: Projectworlds SQL Injection
High CVSS (7.3) vulnerability; prioritize internet-facing assets.
TL;DR - CVE-2026-8785 is a CVSS 7.3 remote SQL-injection flaw in Projectworlds Hospital Management System in PHP 1.0. - The vulnerable
update_info.phpendpoint processes theappointment_noGET parameter; public exploit material exists. - No verified fixed version was identified. Restrict exposure, investigate logs, and replace or securely remediate the application.
Summary
CVE-2026-8785 affects Projectworlds Hospital Management System in PHP version 1.0. The vulnerable functionality is associated with the getAllPatientDetail function in update_info.php, where the application processes the appointment_no HTTP GET parameter. Available reporting describes the issue as remotely exploitable and unauthenticated, but defenders should verify authentication and routing behavior in their own deployment.
For background on related vulnerability terminology, see our guides to CVE-2026-10561 and compensating controls.
| Field | Assessment |
|---|---|
| CVE ID | CVE-2026-8785 |
| Affected product | Projectworlds Hospital Management System in PHP |
| Affected version | 1.0 |
| CVSS base score | 7.3 |
| CVSS vector | Not included in the reviewed NVD API result; retrieve the current NVD record before relying on component-level scoring |
| Attack vector | Network, based on the remote-exploitation description |
| Authentication required | None reported by the linked vulnerability material |
| Vulnerability type | SQL injection, CWE-89 |
| Patch available | No verified vendor patch identified |
| Public exploit | Yes, according to the NVD description and linked GitHub report |
| CISA KEV status | Not listed in the reviewed assessment |
The practical risk depends heavily on deployment. An internet-facing installation may be subject to opportunistic scanning, while an internally restricted installation still presents risk to users or systems that can reach the web application. Because the application handles hospital and patient-related workflows, successful exploitation could expose sensitive records or database credentials, alter data, or provide a foothold for further compromise.
Analyst’s Take: Treat public exposure as the first decision point, not the absence of a CISA KEV listing. The reviewed sources identify a remotely reachable injection point and no verified fixed version, so access restriction and log review should precede any attempt to validate the flaw.
Root Cause
The reported root cause is unsafe handling of the appointment_no input in the getAllPatientDetail function within update_info.php. The linked technical report states that the application uses a custom secure() function in library.php that applies htmlentities(). HTML output encoding does not make a value safe for use in an SQL query. It is designed for a different context and does not replace parameterized database access.
Escaping markup characters or filtering a few obvious payloads does not solve the issue. If attacker-controlled input is concatenated into a SQL statement, the application must use prepared statements with bound parameters. The application should also enforce an appropriate server-side type, such as a positive integer, for appointment_no. Input validation reduces malformed requests but remains a secondary control rather than a substitute for query parameterization.
The reported affected path is:
hospital-management-system-in-php-master/update_info.php
The vulnerability report also describes an access-control weakness around the affected page. Defenders should assess both the SQL construction and whether the endpoint can be reached without an authenticated session. An administrative-looking directory does not provide authorization by itself.
Who’s Exposed
The specifically identified affected product is projectworldsofficial/Hospital-management-system-in-php, version 1.0. The vulnerable component is update_info.php, and the reported input is the GET parameter appointment_no. No additional affected versions were identified in the reviewed sources, so the confirmed version statement is limited to version 1.0 rather than a broader range.
Deployments should be considered exposed when they meet all or most of the following conditions:
| Exposure condition | Why it matters |
|---|---|
| The Projectworlds Hospital Management System in PHP codebase is deployed | The CVE is tied to this application and version |
| The deployed code is version 1.0 or an unverified derivative | No confirmed fixed version was identified |
update_info.php is present and reachable |
This is the reported vulnerable component |
| Requests can reach the endpoint over HTTP or HTTPS | The issue is described as remotely exploitable |
The application accepts attacker-controlled appointment_no input |
This is the reported injection point |
| The database account has broad privileges | Exploitation impact can increase significantly |
There is not enough evidence to claim that every fork, modified copy, or reimplementation is vulnerable. The absence of a formal version banner does not demonstrate safety either. Administrators should compare deployed files and application history against the Projectworlds codebase, identify whether getAllPatientDetail and the relevant query remain present, and test only in an authorized staging environment.
Severity Breakdown
The reported CVSS base score is 7.3, which is generally classified as High. The reviewed NVD API output did not include the complete CVSS vector string, so the precise component values should be retrieved from the current NVD record before publishing a formal scoring analysis or using the score for internal risk calculations.
The available description supports a network-reachable vulnerability with no reported need for local access. The linked GitHub material characterizes the SQL injection as unauthenticated, and the NVD description says the exploit may be performed remotely. These facts increase operational urgency for internet-facing deployments. The final impact depends on database permissions, accessible tables, application behavior, network controls, and whether the vulnerable page is actually exposed.
A practical interpretation of the score is:
- Attack surface: Remote HTTP request handling.
- Privileges: No authentication is reported in the linked technical material.
- User interaction: No user interaction is described as necessary.
- Impact: SQL injection may affect confidentiality, integrity, and availability, depending on the database account and query behavior.
- Environmental modifiers: Internet exposure, sensitive patient data, excessive database privileges, and weak monitoring can make the local risk higher than the base score suggests.
Do not infer an exact CVSS vector from the 7.3 score alone. Retrieve the current NVD record and document any organization-specific environmental scoring separately.
Exploitation Status
A public exploit or proof of concept is available. The NVD description explicitly states that an exploit has been published and may be used. The linked GitHub vulnerability report identifies the endpoint, parameter, and a time-based SQL-injection verification approach. This makes the vulnerability more actionable for attackers than a flaw supported only by a high-level advisory.
Treat the public material as defensive validation content. Security teams should not run time-based injection payloads against production or systems they do not own. If validation is necessary, reproduce the application in an isolated test environment, use synthetic data, and coordinate testing with the system owner.
Confirmed exploitation in the wild has not been established by the reviewed sources. CISA’s Known Exploited Vulnerabilities catalog does not currently list CVE-2026-8785. There is therefore no CISA confirmation, KEV due date, or ransomware-campaign designation for this CVE in the supplied assessment. KEV absence does not prove that exploitation is not occurring; it means only that the CVE lacks that particular public confirmation.
Sources
The primary vulnerability record is the NVD API entry for CVE-2026-8785:
The public technical report and exploit reference are available from the following GitHub repository:
The vulnerability was also reported to the project repository:
Additional vulnerability intelligence references include:
The reviewed assessment found no verified vendor-fixed version and no CISA KEV listing for CVE-2026-8785. Recheck the NVD record, the project repository, and CISA’s catalog before making final publication or remediation decisions because vulnerability status can change after this article’s update date.
This article may contain affiliate links. We earn a commission on qualifying purchases at no extra cost to you.
Detection Guidance
Start with web-server logs, reverse-proxy logs, Web Application Firewall records, and database audit logs. Search for requests to update_info.php that include appointment_no, especially requests containing SQL metacharacters, encoded operators, comments, quotation marks, unusually long values, or timing-related functions. A single suspicious request is not proof of successful exploitation, but repeated variations from one source, followed by elevated database latency or unusual query activity, should be escalated.
A generic access-log pattern can help identify the relevant traffic:
# Examples for Apache, Nginx, or reverse-proxy log review
grep -E '(/|%2f)update_info\.php(\?|%3f).*appointment_no=' access.log
# Review suspicious input characters and common encoded forms
grep -Ei 'update_info\.php.*appointment_no=.*(%27|%22|%23|%2d%2d|'|"|--|/\*|\*/|sleep|benchmark|union|select)' access.log
These patterns are triage aids, not complete signatures. URL encoding, alternate casing, whitespace transformations, HTTP parameter pollution, and application-specific logging formats can evade simple searches. Avoid placing full sensitive query strings in broadly accessible tickets because they may contain patient identifiers or other protected data.
A SIEM query should correlate requests with response and database telemetry. For example, a Splunk-style search might be adapted as follows:
index=web
uri_path="/update_info.php"
uri_query="*appointment_no*"
| eval suspicious=if(match(uri_query,
"(?i)(%27|%22|%23|%2d%2d|union|select|sleep|benchmark|/\*|\*/)"), 1, 0)
| where suspicious=1
| stats count values(http_status) values(user_agent)
values(src_ip) by uri_path, src_ip, bin(_time, 15m)
The exact field names and query syntax must be adjusted to the organization’s logging platform. Investigators should preserve source IP, timestamp, request method, URI, status code, response size, authenticated identity if any, upstream service, and database events. Review for account enumeration, unexpected patient-record queries, data exports, schema changes, new database users, and web-shell or file-write activity after suspicious requests.
Remediation Steps
No verified vendor-issued patch or fixed version was identified in the reviewed references. The confirmed affected version is 1.0, and there is no supported upgrade target that can safely be named. Do not assume that an unrelated fork, a repackaged download, or a changed directory name constitutes a security fix.
The preferred remediation is to remove the application from public exposure and replace it with a maintained system. If the application must remain temporarily available, restrict access through an administrative network or allowlist, and block direct external requests to update_info.php. This reduces exposure but does not correct the underlying vulnerability.
Technical Notes
A temporary web-server restriction can be implemented at the reverse proxy. The following Nginx example denies direct access to the reported endpoint; adapt the location and authentication design to the deployment:
location = /update_info.php {
allow 10.0.0.0/8;
allow 192.168.0.0/16;
allow 172.16.0.0/12;
deny all;
}
If the endpoint is not required, remove or disable it rather than relying solely on network filtering. If it is required, enforce authentication and authorization before processing the request, while recognizing that authentication alone does not remediate SQL injection.
The code-level fix should replace string concatenation with a prepared statement and strict validation. A representative PHP pattern is:
$appointmentNo = filter_input(
INPUT_GET,
'appointment_no',
FILTER_VALIDATE_INT,
['options' => ['min_range' => 1]]
);
if ($appointmentNo === false || $appointmentNo === null) {
http_response_code(400);
exit('Invalid appointment number');
}
$stmt = $pdo->prepare(
'SELECT /* required columns only */
FROM appointments
WHERE appointment_no = :appointment_no'
);
$stmt->execute(['appointment_no' => $appointmentNo]);
The exact table and column names must be taken from the application’s schema. Do not copy this query into production without reviewing the existing business logic. Remove reliance on htmlentities() as an SQL defense, use a database account with only the permissions required by the application, and ensure errors do not disclose SQL statements or credentials.
After containment or code remediation:
- Compare the deployed code with the known Projectworlds 1.0 source and identify all copies of
update_info.php. - Review web, proxy, WAF, and database logs for suspicious
appointment_norequests. - Assume database credentials or patient data may be exposed if exploitation cannot be ruled out.
- Rotate database and application secrets after investigating active use.
- Check for unauthorized database users, modified records, exported data, and persistence.
- Notify privacy, legal, and incident-response teams according to applicable healthcare and breach-reporting requirements.
- Retest the endpoint in an isolated environment using safe, authorized validation methods.
- Add monitoring for future requests to the endpoint and alert on anomalous parameters.