OpenCATS 0.10.0's Career Portal candidate login lacks application-level attempt counters, delays, or rate limiting, allowing remote unauthenticated attackers to repeatedly submit CAPTCHA-checked, low-entropy candidate verification data and potentially disclose candidate information when candidate registration is enabled and publicly reachable. Exploitation is high-complexity (CVSS 4.0 AC:H, PR:N; score 6.3) because each attempt must pass CAPTCHA, and the feature is disabled by default; publicly available exploit code exists, but active exploitation is not confirmed by the supplied KEV data. The issue is a genuine but modest risk rather than an urgent priority, with proxy-layer throttling intended as the practical mitigation.
light0011 CMS hashes user passwords with SHA-1 in App/Home/Model/UserModel.class.php, providing insufficient computational effort and allowing faster offline cracking if stored credential hashes are obtained. Public exploit code exists, but exploitation is difficult and requires prior read access to the user store, typically through a separate weakness such as SQL injection, backup exposure, or database compromise. With confidentiality-only impact, a niche unmaintained codebase, and no confirmed active exploitation, this is a low-real-world-priority issue despite the moderate 6.3 CVSS score.
Resource exhaustion denial of service in AUTOMATIC1111 stable-diffusion-webui up to and including 1.10.1, where a crafted .safetensors model file with a manipulated metadata header-length field causes excessive allocation when the model is loaded. Remote attackers can only trigger this by convincing a user to download and load the malicious model file (mandatory user interaction), so an idle service cannot be attacked directly. Publicly available exploit code exists (a POC zip is referenced), but impact is confined to availability of the web UI process - no data disclosure or code execution - and the project has not released a fix despite being notified via issue report.
Cross-site scripting in AUTOMATIC1111 stable-diffusion-webui (through version 1.10.1) is triggered when the Extra-Networks Card Builder renders the __metadata__.sshs_model_hash field from a safetensors model without output encoding, letting a crafted value execute script in the victim's browser session. Exploitation is unauthenticated from the network perspective but requires the victim to load or import an attacker-supplied model file, so delivery depends on social engineering or a poisoned model repository, and impact is confined to the user's browser rather than leading to server-side code execution. Publicly available exploit code exists (no CISA KEV confirmation), and the project has not responded to the upstream issue, so no patched release is currently identified.
Cross-site scripting in AUTOMATIC1111 stable-diffusion-webui through 1.10.1 lets an attacker inject script into the Extra Networks 'Card Builder' view by crafting a malicious filename for a model, LoRA, embedding, or hypernetwork card that create_item_html in modules/ui_extra_networks.py renders without proper escaping. Attackers must first place a file with that crafted name into a scanned networks directory; a user then triggers execution simply by opening the affected view, so no credentials are needed (CVSS 4.0 base 5.3, PR:N/UI:P). Publicly available exploit code exists (E:P) and the vendor has not responded to the issue report, but impact is limited to low-integrity script execution in the victim's browser, and the default localhost bind means remote exploitation requires a deliberately network-exposed instance (--listen or --share).
Code injection in AUTOMATIC1111 Stable Diffusion WebUI through 1.10.1 lets remote attackers execute attacker-chosen Python objects when a victim loads a malicious third-party model whose YAML configuration supplies a crafted 'target' value to get_obj_from_str in modules/sd_models.py. User interaction is mandatory - the victim must import the untrusted model/config - but no authentication is needed on the attacker side because the payload travels inside distributed model files; publicly available exploit code exists and a proof-of-concept ZIP is referenced. No CISA KEV entry confirms active exploitation, and the vendor has not responded to the early issue report, so no vendor-released patch was identified at time of analysis; the supplied CVSS 4.0 score of 2.1 conflicts with the independently assessed CVSS 3.1 vector, which rates post-load confidentiality, integrity, and availability impact as high.
SQL injection in the Exam Module of itsmeravitejak Online Examination System allows an authenticated low-privilege user to manipulate the topicid argument in exam.php's Exam::startExam function and alter database queries, with the independent assessment indicating high confidentiality impact, low integrity impact, and no availability impact. The issue affects the unversioned rolling-release codebase through commit e6208e630b14657999bbfe4c9912302d2a1e2449, and publicly available exploit code exists; there is no CISA KEV confirmation of active exploitation. Despite the public exploit, real-world priority is low because a valid application login is required and no vendor-fixed release is available.
Reflected cross-site scripting (CWE-79) in the Search argument of /abs.php in hrisikeshpal Blood Bank Management System 1.0.0 lets a remote, unauthenticated attacker inject script that executes in a victim's browser session. Exploitation requires the victim to load an attacker-crafted request (UI:P passive user interaction), so it depends on link-delivery or social engineering against a running instance rather than direct compromise; confidentiality and availability impact is assessed as none, with only limited integrity impact inside the victim's session. Publicly available exploit code exists and the flaw was disclosed through a GitHub issue that the project has not yet answered; no vendor-released patch was identified at time of analysis, and EPSS/KEV data was not provided.
Java deserialization in DataStax pulsar-jms through 9.0.3 allows a low-privileged publisher to trigger unsafe object decoding in consuming JMS clients via crafted ObjectMessage content, potentially causing limited confidentiality, integrity, and availability impact, or code execution when a suitable gadget chain is present. Publicly available exploit code exists, but exploitation requires the consumer to process ObjectMessage payloads and depends on classpath gadgets for meaningful code-execution impact. No vendor-released fixed version is identified, and the project has not responded to the reported issue.
Default-password exposure in Diem CMS through 5.1.3 allows remote attackers to access unhardened installations whose credential, provisioned through dmCorePlugin/lib/data/dmDataLoad.php, was never rotated. Publicly available exploit code exists and the vendor has not responded to the project issue; the practical risk is limited to deployments with the default configuration left in place, making this a low-to-moderate priority issue despite network reachability.
Missing authorization in the Scheduled Restore component of David-Crty Databasement 1.8.1 and earlier lets an authenticated user supply an arbitrary schema_name value to the SaveScheduledRestoreRequest API endpoint and thereby schedule restore operations against database schemas they are not authorized to manage. Exploitation is network-reachable but requires a valid privileged session on the instance plus the Scheduled Restore feature being in use, and public exploit code exists per the vendor GitHub security advisory. The issue is fixed in release 1.8.2; impact per the assessed vector is limited to low confidentiality, integrity and availability effects on the appliance's own scope.
Cleartext password storage in Soarkey StudentManagement through commit e08f7f1d5015af407aa4cca0ada3dea189b4937e allows local low-privileged users to recover sensitive credentials handled by the UserDao.login flow in code/src/service/AdminDao.java. The issue is CWE-312 and is local-only with a CVSS score of 1.9; publicly available exploit code exists, but the project is an unmaintained student codebase with negligible production deployment, so real-world risk is very low.
Weak SHA1-based password hashing in the Diem PHP framework's PluginDmUser handler (lib/model/doctrine/PluginDmUser.class.php, versions through 5.1.3) lets anyone who already holds the stored password hashes from a Diem user database recover plaintext credentials offline far faster than a modern key-derivation function would allow. Exploitation is strictly local and high-complexity and presupposes a prior foothold to obtain the hashes, so the weakness yields no remote or unauthenticated path into the application; impact is limited to confidentiality of credentials and any password reuse elsewhere. Publicly available exploit code exists (referenced upstream issue 452), the project was notified but has not responded, and no CISA KEV listing or EPSS score was supplied with the intelligence.
Incorrect authorization in Gitea's label-deletion logic lets an authenticated user with write access to a repository's issues delete issue labels belonging to an unrelated organization whose numeric ID happens to match that repository's ID. Exploitation requires the coincidental ID collision between a repository and an organization, cannot be chosen by the attacker, and yields only limited integrity impact (deleted labels) with no confidentiality or availability loss; Gitea instances through 28.1.0 are affected and 28.2.0 fixes the flaw. Risk is low: EPSS is 0.15% (4th percentile), SSVC marks exploitation as none and non-automatable, and no public exploit code was identified at time of analysis.
Established BGP peers can bypass IPv6 FlowSpec validation in CZ.NIC BIRD when the receiving flow6 channel has unicast-route validation enabled, because BIRD reconstructs a prefix with zero-filled skipped bits instead of rejecting a destination component with a nonzero offset as RFC 8956 section 5 requires. This allows the peer to inject a FlowSpec rule matching a broader set of IPv6 destinations than its advertised unicast route authorizes, potentially causing downstream traffic filtering or redirection where export policy permits propagation and downstream devices enforce the rule. Exploitation is limited to authenticated peers with a suitable unicast route and the validation-enabled flow6 configuration; no public exploit was identified at time of analysis, and an upstream fix is available via commit 7cf75587b6f38b87849fe192bf5dfd124b7702bc.
Authenticated users of Teable can rename folders that belong to bases they have no membership in by submitting a folderId from a foreign base to the PATCH /api/base/:baseId/node/folder/:folderId endpoint, which validates the base|update permission only against the baseId in the URL rather than against the base that actually owns the folder. The flaw is an IDOR / authorization bypass (CWE-639) affecting releases prior to release.2026-10-12T12-45-53Z.3331; the vendor-supplied CVSS 4.0 vector scores it 2.3 and our independent CVSS 3.1 assessment scores it AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N, reflecting integrity-only impact with no data disclosure or destruction. Exploitation requires a valid Teable session plus discovery of a non-enumerable foreign folderId, so multi-tenant or shared deployments carry the most practical exposure; no public exploit code was identified at time of analysis and there is no confirmed active exploitation (not in CISA KEV).
Gitify desktop builds prior to 7.10.0 perform the GitHub web OAuth flow without generating or validating an OAuth `state` value, so an attacker who supplies their own authorization code via a crafted `gitify://oauth?code=...` deep link can bind a victim's Gitify client to the attacker's GitHub account. The attacker then controls the notifications the victim sees, but gains no access to the victim's data and no account takeover - impact is limited to low integrity. Exploitation is unauthenticated (PR:N) but requires a race against an already-pending login (AT:P) and victim interaction with the attacker's link (UI:P). No public exploit code or CISA KEV entry was identified at time of analysis, and the issue is assessed as genuinely low priority.
Code injection in WongKinYiu YOLOv9 0.1 via the ONNX metadata handler allows an attacker who supplies a tampered .onnx model to execute code when DetectMultiBackend parses the model's 'names' and 'stride' fields; exploitation is unauthenticated but requires the victim to load the untrusted model (UI:P), making the realistic threat a supply-chain or malicious-pretrained-model scenario rather than direct compromise of a passive service. Publicly available exploit code exists, and the project has not responded to the issue report; signals are mixed and point to a low operational priority despite a serious weakness class.
Reflected cross-site scripting (CWE-79) in the abandoned open-source PHP application Books4Cash allows unauthenticated remote attackers to inject arbitrary script into the search page by crafting a malicious 'Search' argument accepted by search.php. The payload is returned to the browser without sanitization, so the impact is confined to the victim's session - an attacker can execute script in the context of the site, deface or alter content, and potentially capture session data or perform actions as the victim. Exploitation is not unattended: a victim must follow an attacker-supplied link (CVSS UI:P social-engineering prerequisite), and publicly available exploit code exists per the CVSS 4.0 E:P metric and VulDB disclosure. Real-world exposure is limited because the project repository has been dormant for years and has a negligible deployment footprint; no CISA KEV entry indicates active exploitation.
Stored cross-site scripting in Chamilo LMS's Ticket Module lets authenticated users who can create or edit ticket projects inject a script payload through the project Name field, which executes when another user views the affected page. The flaw affects Chamilo LMS up to 1.11.40, 2.0.3, and 3.0.1 and requires victim interaction; publicly available exploit code has been disclosed, though no CISA KEV active-exploitation confirmation is present. The vendor fix is visible in PR #9228 and commit cb6fe2127e587603451c609b4ed78c0293fcc3fa, and builds applying that output-encoding patch are not exploitable.
Terminal escape sequence injection in alexpasmantier television through 0.15.9 allows a local attacker who controls an entry's output, such as a crafted filename, to write unsanitized control characters to the victim's terminal after the TUI exits, provided the victim selects that entry. The CVE is tracked as CWE-77 command injection and rated CVSS 4.8, but PR #1125 adds sanitize_for_display for control characters, indicating the practical issue is terminal escape/control-character injection rather than direct command execution. The CVSS vector notes E:P (Proof-of-Concept exploit maturity), but no CISA KEV active exploitation is confirmed; exploitation requires local staging and victim interaction, making this a low-priority issue.