Writeup Exploits
51,780 exploits tracked across all sources.
Erlang OTP ssh_sftpd - Path Traversal
Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') vulnerability in Erlang OTP (ssh_sftpd module) allows Path Traversal.
This vulnerability is associated with program files lib/ssh/src/ssh_sftpd.erl and program routines ssh_sftpd:is_within_root/2.
The SFTP server uses string prefix matching via lists:prefix/2 rather than proper path component validation when checking if a path is within the configured root directory. This allows authenticated users to access sibling directories that share a common name prefix with the configured root directory. For example, if root is set to /home/user1, paths like /home/user10 or /home/user1_backup would incorrectly be considered within the root.
This issue affects OTP from OTP 17.0 until OTP 28.4.1, OTP 27.3.4.9 and OTP 26.2.5.18, corresponding to ssh from 3.0.1 until 5.5.1, 5.2.11.6 and 5.1.4.14.
Erlang OTP ssh - DoS
Improper Handling of Highly Compressed Data (Compression Bomb) vulnerability in Erlang OTP ssh (ssh_transport modules) allows Denial of Service via Resource Depletion.
The SSH transport layer advertises legacy zlib compression by default and inflates attacker-controlled payloads pre-authentication without any size limit, enabling reliable memory exhaustion DoS.
Two compression algorithms are affected:
* zlib: Activates immediately after key exchange, enabling unauthenticated attacks
* [email protected]: Activates post-authentication, enabling authenticated attacks
Each SSH packet can decompress ~255 MB from 256 KB of wire data (1029:1 amplification ratio). Multiple packets can rapidly exhaust available memory, causing OOM kills in memory-constrained environments.
This vulnerability is associated with program files lib/ssh/src/ssh_transport.erl and program routines ssh_transport:decompress/2, ssh_transport:handle_packet_part/4.
This issue affects OTP from OTP 17.0 until OTP 28.4.1, 27.3.4.9 and 26.2.5.18 corresponding to ssh from 3.0.1 until 5.5.1, 5.2.11.6 and 5.1.4.14.
Incomplete control character validation in http.cookies
The fix for CVE-2026-0672, which rejected control characters in http.cookies.Morsel, was incomplete. The Morsel.update(), |= operator, and unpickling paths were not patched, allowing control characters to bypass input validation. Additionally, BaseCookie.js_output() lacked the output validation applied to BaseCookie.output().
Stack overflow parsing XML with deeply nested DTD content models
When an Expat parser with a registered ElementDeclHandler parses an inline
document type definition containing a deeply nested content model a C stack
overflow occurs.
mackron / dr_libs Excessive Memory Allocation in PICTURE Metadata Parsing
dr_libs dr_flac.h version 0.13.3 and earlier (fixed in commits fefced4, 4f5a4cd, and 663239a) contain an uncontrolled memory allocation vulnerability in drflac__read_and_decode_metadata() that allows attackers to trigger excessive memory allocation by supplying crafted PICTURE metadata blocks. Attackers can exploit attacker-controlled mimeLength and descriptionLength fields to cause denial of service through memory exhaustion when processing FLAC streams with metadata callbacks.
CVSS 6.2
OpenClaw < 2026.2.22 BlueBubbles - Access Control Bypass via Empty allowFrom Configuration
OpenClaw versions prior to 2026.2.22 with the optional BlueBubbles plugin contain an access control bypass vulnerability where empty allowFrom configuration causes dmPolicy pairing and allowlist restrictions to be ineffective. Remote attackers can send direct messages to BlueBubbles accounts by exploiting the misconfigured allowlist validation logic to bypass intended sender authorization checks.
CVSS 6.5
OpenClaw < 2026.2.19 - Path Traversal in Feishu Media Temporary File Naming
OpenClaw versions prior to 2026.2.19 contain a path traversal vulnerability in the Feishu media download flow where untrusted media keys are interpolated directly into temporary file paths in extensions/feishu/src/media.ts. An attacker who can control Feishu media key values returned to the client can use traversal segments to escape os.tmpdir() and write arbitrary files within the OpenClaw process permissions.
CVSS 8.2
OpenClaw < 2026.2.26 - Approval Bypass via Parent Symlink Current Working Directory Rebind
OpenClaw versions prior to 2026.2.26 contain an approval bypass vulnerability in system.run execution that allows attackers to execute commands from unintended filesystem locations by rebinding writable parent symlinks in the current working directory after approval. An attacker can modify mutable parent symlink path components between approval and execution time to redirect command execution to a different location while preserving the visible working directory string.
CVSS 6.1
OmniGen2-RL Reward Server Unsafe Deserialization RCE
OmniGen2-RL contains an unauthenticated remote code execution vulnerability in the reward server component that allows remote attackers to execute arbitrary commands by sending malicious HTTP POST requests. Attackers can exploit insecure pickle deserialization of request bodies to achieve code execution on the host system running the exposed service.
CVSS 9.8
HTSlib CRAM reader has out-of-bounds read due to improper validation of input
HTSlib is a library for reading and writing bioinformatics file formats. CRAM is a compressed format which stores DNA sequence alignment data. As one method of removing redundant data, CRAM uses reference-based compression so that instead of storing the full sequence for each alignment record it stores a location in an external reference sequence along with a list of differences to the reference at that location as a sequence of "features". When decoding CRAM records, the reference data is stored in a char array, and parts matching the alignment record sequence are copied over as necessary. Due to insufficient validation of the feature data series, it was possible to make the `cram_decode_seq()` function copy data from either before the start, or after the end of the stored reference either into the buffer used to store the output sequence for the cram record, or into the buffer used to build the SAM `MD` tag. This allowed arbitrary data to be leaked to the calling function. This bug may allow information about program state to be leaked. It may also cause a program crash through an attempt to access invalid memory. Versions 1.23.1, 1.22.2 and 1.21.1 include fixes for this issue. There is no workaround for this issue.
CVSS 9.1
pkgutil.get_data() does not enforce documented restrictions
DISPUTED: The project has clarified that the documentation was incorrect, and that pkgutil.get_data() has the same security model as open(). The documentation has been updated to clarify this point. There is no vulnerability in the function if following the intended security model.
pkgutil.get_data() did not validate the resource argument as documented, allowing path traversals.
Discourse has check revision visibility on posts endpoint
Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, requesting /posts/:id.json?version=X bypassed authorization checks on post revisions. The display_post method called post.revert_to directly without verifying whether the revision was hidden or if the user had permission to view edit history. This meant hidden revisions (intentionally concealed by staff) could be read by any user by simply enumerating version numbers. Starting in versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, Discourse looks up the PostRevision and call guardian.ensure_can_see! before reverting, consistent with how the /posts/:id/revisions/:revision endpoint already authorizes access. No known workarounds are available.
CVSS 5.3
Discourse has a bypass of official warnings messages by non-staff users
Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, a type coercion issue in a post actions API endpoint allowed non-staff users to issue warnings to other users. Warnings are a staff-only moderation feature. The vulnerability required the attacker to be a logged-in user and to send a specifically crafted request. No data exposure or privilege escalation beyond the ability to create unauthorized user warnings was possible. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.
CVSS 4.3
Discourse Vulnerable to Stored XSS via Shared AI Conversation Onebox
Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, the onebox method in the SharedAiConversation model renders the conversation title directly into HTML without proper sanitization. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. As a workaround, tighten access by changing the `ai_bot_public_sharing_allowed_groups` site setting.
CVSS 6.1
Discourse has Stored XSS in AI Triage Automation
Discourse is an open-source discussion platform. Versions prior to 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 have a cross-site scripting vulnerability that arises because the system trusts the raw output from an AI Large Language Model (LLM) and renders it using htmlSafe in the Review Queue interface without adequate sanitization. A malicious attacker can use valid Prompt Injection techniques to force the AI to return a malicious payload (e.g., tags). When a Staff member (Admin/Moderator) views the flagged post in the Review Queue, the payload executes. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. As a workaround, temporarily disable AI triage automation scripts.
CVSS 6.1
Discourse leaks private topic metadata to non-authorized users
Discourse is an open-source discussion platform. Versions prior to 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 have a vulnerability in an API endpoint that discloses private topic metadata of admin users to moderator users even if the moderators do not have access to the private topics. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.
ormar has a Pydantic Validation Bypass via Kwargs Injection in Model Constructor
ormar is a async mini ORM for Python. Versions 0.23.0 and below are vulnerable to Pydantic validation bypass through the model constructor, allowing any unauthenticated user to skip all field validation by injecting "__pk_only__": true into a JSON request body. By injecting "__pk_only__": true into a JSON request body, an unauthenticated attacker can skip all field validation and persist unvalidated data directly to the database. A secondary __excluded__ parameter injection uses the same pattern to selectively nullify arbitrary model fields (e.g., email or role) during construction. This affects ormar's canonical FastAPI integration pattern recommended in its official documentation, enabling privilege escalation, data integrity violations, and business logic bypass in any application using ormar.Model directly as a request body parameter. This issue has been fixed in version 0.23.1.
CVSS 7.1
Discourse vulnerable to group membership addition permission bypass via discourse-policy plugin
Discourse is an open-source discussion platform. Versions prior to 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 have a security flaw in the discourse-policy plugin which allowed a user with policy creation permission to gain membership access to any private/restricted groups. Once membership to a private/restricted group has been obtained, the user will be able to read private topics that only the group has access to. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. As a workaround, review all policies for the use of `add-users-to-group` and temporarily remove the attribute from the policy. Alternatively, disable the discourse-policy plugin by disabling the `policy_enabled` site setting.
OpenClaw < 2026.3.2 - Authentication Bypass via Encoded Path in /api/channels Route
OpenClaw versions prior to 2026.3.2 contain an authentication bypass vulnerability in the /api/channels route classification due to canonicalization depth mismatch between auth-path classification and route-path canonicalization. Attackers can bypass plugin route authentication checks by submitting deeply encoded slash variants such as multi-encoded %2f to access protected /api/channels endpoints.
CVSS 6.5
OpenClaw < 2026.2.19 - Arbitrary File Write via Short-Option Bypass in exec Allowlist
OpenClaw versions prior to 2026.2.19 contain an allowlist bypass vulnerability in the exec safeBins policy that allows attackers to write arbitrary files using short-option payloads. Attackers can bypass argument validation by attaching short options like -o to whitelisted binaries, enabling unauthorized file-write operations that should be denied by safeBins checks.
CVSS 7.1
OpenClaw < 2026.2.22 - Incomplete IPv4 Special-Use Range Blocking in SSRF Guard
OpenClaw versions prior to 2026.2.22 contain incomplete IPv4 special-use range validation in the isPrivateIpv4() function, allowing requests to RFC-reserved ranges to bypass SSRF policy checks. Attackers with network reachability to special-use IPv4 ranges can exploit web_fetch functionality to access blocked addresses such as 198.18.0.0/15 and other non-global ranges.
CVSS 7.4
OpenClaw < 2026.2.24 - Arbitrary File Read via Improper Temporary Path Validation in Sandbox
OpenClaw versions prior to 2026.2.24 contain an improper path validation vulnerability in sandbox media handling that allows absolute paths under the host temporary directory outside the active sandbox root. Attackers can exploit this by providing malicious media references to read and exfiltrate arbitrary files from the host temporary directory through attachment delivery mechanisms.
CVSS 6.5
Discourse filters whisper posts from private-posts feed
Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, the `/private-posts` endpoint did not apply post-type visibility filtering, allowing regular PM participants to see whisper posts in PM topics they had access to. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.
CVSS 6.5
Discourse fixes loose hostname matching in spam host allowlist
Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, the `allowed_spam_host_domains` check used `String#end_with?` without domain boundary validation, allowing domains like `attacker-example.com` to bypass spam protection when `example.com` was allowlisted. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 require exact match or proper subdomain match (preceded by `.`) to prevent suffix-based bypass of `newuser_spam_host_threshold`. No known workarounds are available.
CVSS 4.3
Discourse leaks PM post edits to moderators
Discourse is an open-source discussion platform. Prior to versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2, the Post Edits admin report (/admin/reports/post_edits) leaked the first 40 characters of raw post content from private messages and secure categories to moderators who shouldn't have access. Versions 2026.3.0-latest.1, 2026.2.1, and 2026.1.2 contain a patch. No known workarounds are available.
CVSS 2.7
By Source