CVE Vulnerability Catalog

MCP Atlassian prior to 0.22.0 writes a plaintext fallback file containing access and refresh tokens under the user's .mcp-atlassian directory using process-default permissions. On systems with a permissive umask, same-group or other local users and processes can read the persisted tokens and reuse the associated Atlassian access. This issue is fixed in version 0.22.0.

RTI Connext Professional (Core Libraries) has an Out-of-bounds Read vulnerability that allows overread buffers. This affects versions from 7.4.0 before 7.7.0.1, from 7.0.0 before 7.3.1.6, from 6.1.0 before 6.1.*, from 6.0.0 before 6.0.*, from 5.3.0 before 5.3.*, from 5.2.0 before 5.2.*, from 5.0.0 before 5.1.*.

RTI Connext Professional (Core Libraries) has an Off-by-one Error and Out-of-bounds Write vulnerability that allows overflow buffers. This affects versions from 7.4.0 before 7.7.0.1 and from 7.0.0 before 7.3.1.6.

RTI Connext Professional (Core Libraries) has an Out-of-bounds Read, Function Call With Incorrect Number of Arguments, and Access of Resource Using Incompatible Type ('Type Confusion') vulnerability that allows overread buffers. This affects versions from 7.4.0 before 7.7.0.1, from 7.3.0 before 7.3.1.6, and from 6.1.2.21 before 6.1.*.

RTI Connext Professional (Core Libraries) has an Out-of-bounds Read, Function Call With Incorrect Number of Arguments, and Access of Resource Using Incompatible Type ('Type Confusion') vulnerability that allows overread buffers. This affects versions from 7.4.0 before 7.7.0.1 and from 7.3.0 before 7.3.1.6.

RTI Connext Professional (Core Libraries) has a Double Free vulnerability that allows file manipulation. This affects versions from 7.4.0 before 7.7.0.1 and from 7.1.0 before 7.3.1.6.

LightRAG prior to 1.5.5 renders answer and thinking content with react-markdown, rehypeRaw, and skipHtml=false without an HTML sanitizer. An attacker who can add a document can store raw HTML that is returned through the query path and rendered as active content, enabling arbitrary JavaScript execution via elements such as iframe srcdoc. The script runs in the LightRAG WebUI origin and can read the API token from localStorage and perform API actions as the victim.

LightRAG prior to 1.5.5 has verify_password in lightrag/api/passwords.py comparing plaintext AUTH_ACCOUNTS password values with Python's == operator. The comparison can return after the first mismatching byte, creating response-time differences based on password length and matching prefixes. A network attacker with sufficiently low-latency access can repeatedly query /login and use the timing oracle to recover a plaintext-configured password character by character. Deployments using bcrypt-prefixed password values are not affected by this comparison path.

LightRAG prior to 1.5.5 returns raw Python exception text from error handlers in document_routes.py, graph_routes.py, query_routes.py, ollama_api.py, and lightrag_server.py. The detail=str(e), detail=str(exc), and equivalent formatted-message paths expose server filesystem paths, database host, port, user, and database names, language-model provider diagnostics, configuration details, and Python library internals to a network client that can trigger an error. The default unauthenticated configuration makes those responses reachable without credentials.

FastGPT prior to 4.15.2 has a DNS rebinding flaw in the safe Axios request interceptor in packages/service/common/api/axios.ts. The hostname is validated with isInternalAddress() before a later HTTP connection performs an independent DNS lookup, creating a time-of-check/time-of-use gap. An authenticated attacker who can supply a URL to a safe-Axios-backed HTTP tool, workflow HTTP node, external file fetch, or other server-side integration can reach services that direct private-address inputs would block.

In Netdata from 2.0.0 until 2.11.0, the /api/v3/settings endpoint is registered with HTTP_ACL_NOCHECK and HTTP_ACCESS_ANONYMOUS_DATA, allowing unauthenticated PUT requests to bypass operator-configured IP allow restrictions. An attacker can persist attacker-controlled JSON in {varlib}/settings/default.json, manipulate its version counter, and use repeated near-20 MiB writes to consume disk space.

In Netdata prior to 2.10.4, an authenticated child agent can send an oversized DIMENSION SLOT value that str2ull_encoded passes to pluginsd_rrddim_put_to_slot without an upper bound. prd_array_create can then wrap the size_t allocation calculation while retaining the original large array size, causing the initialization loop to write beyond the undersized heap allocation and crash the parent agent.

In Netdata prior to 2.10.4, an authenticated child agent can send an oversized CHART SLOT value that str2ull_encoded passes to pluginsd_rrdset_cache_put_to_slot. The accepted slot drives reallocz to request an approximately 16 GiB chart-pointer array, and allocation failure invokes fatal and aborts the parent Netdata agent.

In Nuclei from 3.0.0 until 3.10.0, the DAST/fuzz payload path in pkg/fuzz/parts.go can evaluate substituted runtime data more than once, creating a second evaluation pass that allows response content captured by an internal extractor to be reinterpreted as template syntax. When -env-vars is enabled, a malicious scan target can return marker-shaped content that discloses scanner-host environment variables such as credentials, API keys, or tokens.

In Nuclei from 3.0.0 until 3.10.0, the workflow template loading path does not enforce the -file capability gate when resolving file: protocol templates referenced by a workflow. An untrusted unsigned workflow can therefore load a file-protocol template and read local files from the scanner host even when the operator has not enabled file templates.

In Nuclei from 3.0.0 until 3.10.0, the nuclei/mysql JavaScript library does not enforce the local-file sandbox when a JavaScript template supplies the allowAllFiles MySQL DSN option. An untrusted javascript: template scanning an attacker-controlled MySQL-compatible endpoint can cause the server to issue LOAD DATA LOCAL INFILE requests for arbitrary paths, after which the scanner's MySQL client reads the files and returns their contents to the server even though -allow-local-file-access is disabled.

In Nuclei from 3.0.0 until 3.10.0, the DAST template loading branch does not apply the unsigned code-template signature check before accepting a template that contains both a fuzzing: block and an unsigned code: block. When an operator enables -dast, an untrusted multiprotocol template can place an unsigned code request into the execution queue and run arbitrary shell commands even without -code or a valid cryptographic signature.

In 9Router prior to 0.5.6, deployments that allow requests to reach Next.js without the sanitizing custom-server.js wrapper use the client-supplied X-9r-Real-Ip value as the bucket key in getClientIp, checkLock, and recordFail in src/lib/auth/loginLimiter.js for POST /api/auth/login. A remote unauthenticated attacker can rotate the header on every password guess so each request uses a new failed-attempt bucket and the five-attempt progressive lockout never returns HTTP 429.

A typo in the MISP ACLComponent access control configuration caused the ACL rule for the previewEventAttributes action to reference the permission string 'theming_enabled*' (with a trailing asterisk) instead of the correct 'theming_enabled'. The malformed key does not match any valid permission identifier, causing the access control check for previewEventAttributes to malfunction — potentially resulting in either an authorization bypass or denial of access for legitimate users.

In SSSD configured with the LDAP access provider and ldap_access_order including ppolicy or lockout, a fail-open condition in the LDAP ppolicy access check can occur if a user lookup returns zero results. This can incorrectly return success and cache an allow decision, permitting continued authorization for a deleted or deprovisioned user.