🚨 CVE-2026-64679
Atlantis is a self-hosted golang application that listens for Terraform pull request events via webhooks. From 0.19.8 until 0.45.0, Atlantis does not consistently validate user-controlled workspace values supplied through accepted repository-level atlantis.yaml configuration or authenticated /api/plan input before joining them into local workspace paths. Traversal segments can escape the intended per-pull workspace directory and cause clone preparation or other working-directory code paths to call os.RemoveAll, os.MkdirAll, or related filesystem operations on out-of-bounds directories before Terraform rejects the invalid workspace name. This can create, delete, or reuse writable paths with the privileges of the Atlantis process, causing integrity loss or denial of service. This issue is fixed in version 0.45.0.

🎖@cveNotify

🚨 CVE-2026-77415
JSONata is a JSON query and transformation language. Prior to 1.8.8 and 2.2.1, crafted JSONata expressions could chain several object-integrity weaknesses to execute arbitrary code. The chain could overwrite $clone to mutate objects through evaluateTransformExpression, expose and deconstruct JSONata functions or lambdas through $merge.*, replace proc.arguments.forEach used by applyProcedure, and forge internal lambda state. These primitives allowed an attacker to reach prototype getters, prototype and constructor access, and process.getBuiltinModule with child_process, executing code with the privileges of the host process. This issue is fixed in versions 1.8.8 and 2.2.1.

🎖@cveNotify

🚨 CVE-2026-49849
xShop is an open-source shop developed in Laravel. An Unrestricted File Upload vulnerability in xShop version 3.0.3 allows an authenticated administrator to upload executable files (e.g., .php). By uploading a specially crafted php file, an attacker can achieve Remote Code Execution (RCE) on the server, leading to a full system compromise. Version 3.0.4 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-53487
Kite is a Kubernetes dashboard. Prior to version 0.12.3, authenticated Kite users with any role can request `/api/v1/overview` for a cluster that their roles do not permit by selecting that cluster with `x-cluster-name`. The overview route is registered before `middleware.RBACMiddleware()` and `GetOverview` only checks `len(user.Roles) > 0`, so it returns aggregate Kubernetes inventory and capacity data from unauthorized clusters. Version 0.12.3 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-53531
RaTeX is a KaTeX-compatible math rendering engine written in Rust. Prior to version 0.1.11, RaTeX’s recursive-descent parser recurses one (or more) native stack frame per nesting level at `{`, `\left`, `\sqrt{`, `^{`, etc, with no maximum depth limit. A short, ~10 KB input of nested groups overflows the 8 MB main-thread stack and aborts the process. With `panic = "abort"` (`Cargo.toml:48`), and because a Rust stack overflow is always a fatal `SIGABRT` regardless of panic strategy this is an unrecoverable, whole-process denial of service reachable from a single untrusted LaTeX string. Version 0.1.11 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-48050
Arc is an open, SQL-native time-series database for telemetry. Versions prior to 26.06.1 register Go's `net/http/pprof` handlers at `/debug/pprof/*` via `app.Use(pprof.New())` in `internal/api/server.go`, and `/debug/pprof` is added to `PublicPrefixes` in `cmd/arc/main.go`. The auth middleware short-circuits before the token check on prefix match, so the endpoints are reachable without any authentication. Version 26.06.1 contains a patch. Some workarounds are available. Block `/debug/pprof*` at a reverse proxy / load balancer in front of Arc, restrict Arc's API port to known-trusted networks via firewall rules, and/or patch the running build: comment out `app.Use(pprof.New())` in `internal/api/server.go` and rebuild.

🎖@cveNotify

🚨 CVE-2026-53525
WeeChat (Wee Enhanced Environment for Chat) is a free chat client. In versions 0.3.1 through 4.9.0, the WeeChat relay authentication uses non-constant-time string comparison functions (weechat_strcasecmp and strcmp) to verify password hashes and plaintext passwords. An attacker can exploit timing differences to extract the server-computed hash character by character, then authenticate using the correct hash without knowing the password. Version 4.9.1 fixes the issue.

🎖@cveNotify

🚨 CVE-2026-57998
better-npm-audit through 3.11.0, and the 4.0.0-rc.2 prerelease, builds its npm audit command by interpolating the user-supplied --registry option into a command string in src/handlers/handleInput.ts without validation or quoting, then passes that string to child_process.exec() in index.ts, which spawns a shell. A registry value containing shell metacharacters such as a semicolon, pipe, or command substitution executes arbitrary operating system commands with the privileges of the process running the audit.

🎖@cveNotify

🚨 CVE-2026-72704
The guard checker in Rocq Prover does not recheck the recursive tree representation of an inductive type parameter after that parameter has been changed by transport. A fixpoint may apply a rewrite along an equality between types to its recursive argument, which the guard checker accepts because the inductive type is preserved, while the recursive tree recorded for the parameter is altered. A second fixpoint that calls the first inherits the altered recursive tree without verification, so a call that is not structurally decreasing is accepted as terminating. The resulting non-terminating definition proves that a natural number equals its own successor and therefore False, from which any proposition follows. The demonstration uses two axioms that follow from univalence and are consistent with the calculus of inductive constructions, so the contradiction comes from the guard check rather than from the assumptions. A fix is proposed but not merged.

🎖@cveNotify

🚨 CVE-2026-75509
joserfc is a Python library that provides an implementation of several JSON Object Signing and Encryption (JOSE) standards. Prior to version 1.7.3, JWTClaimsRegistry applies membership matching to list-valued iss and sub claims, allowing an array-valued iss that contains the expected issuer to pass an intended equality check and enabling issuer-validation bypass. This issue is fixed in version 1.7.3.

🎖@cveNotify

🚨 CVE-2026-55525
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, the web_crawl function validates only the initial URL before _crawl_with_httpx uses httpx.Client(follow_redirects=True). Redirect targets are not revalidated, so an attacker who influences a crawl target can redirect a public URL to loopback, private network, or cloud metadata services while ALLOW_LOCAL_CRAWL remains disabled. The fetched internal response is returned to the agent context. This issue is fixed in version 1.6.58.

🎖@cveNotify

🚨 CVE-2026-55530
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, ast_grep_rewrite lacks the @require_approval decorator used by sibling mutation tools. With dry_run=False, an agent-controlled call can pass --update-all and a broad path to rewrite matching files without the expected authorization gate. This issue is fixed in version 1.6.58.

🎖@cveNotify

🚨 CVE-2026-55531
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, the MCP HTTP Stream mcp_post handler creates a new _sessions entry for every initialize request but does not call _cleanup_sessions or enforce a maximum. An unauthenticated caller can exhaust memory. The fix invokes cleanup and limits sessions through PRAISONAI_MCP_MAX_SESSIONS. This issue is fixed in version 4.6.58.

🎖@cveNotify

🚨 CVE-2026-55537
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, JobSubmitRequest.validate_webhook_url() accepts webhook_url when resolution raises socket.gaierror because the exception path uses except socket.gaierror: pass. JobExecutor._send_webhook() later performs a fresh lookup, allowing DNS changes to direct the request to an internal service. This issue is fixed in version 4.6.58.

🎖@cveNotify

🚨 CVE-2026-55538
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.51, praisonai serve agents parses config["api_key"] but _create_agents_app() does not authenticate POST /agents or POST /agents/{agent_name}. Missing or incorrect bearer and X-API-Key values still reach agent execution. This issue is fixed in version 4.6.58.

🎖@cveNotify

🚨 CVE-2022-50999
Nokogiri versions before 1.13.5 contain an integer overflow vulnerability in packaged libxml2 buffer handling functions that allows attackers to cause out-of-bounds memory writes. Attackers can exploit this by crafting multi-gigabyte XML files to trigger buffer overflows resulting in information disclosure, data modification, or denial of service.

🎖@cveNotify

🚨 CVE-2023-54354
Nokogiri before 1.14.3 (CRuby implementation only, when using the packaged libxml2) bundles libxml2 v2.10.3, which is vulnerable to NULL pointer dereferences in XML Schema processing (xmlSchemaFixupComplexType, CVE-2023-28484, and xmlSchemaCheckCOSSTDerivedOK). An attacker who supplies a crafted/malformed XML schema can cause libxml2 to dereference a NULL pointer and potentially segfault, resulting in a denial of service. Nokogiri 1.14.3 upgrades the packaged libxml2 to v2.10.4 to resolve these issues.

🎖@cveNotify

🚨 CVE-2025-71346
Nokogiri before 1.18.8 packages a vulnerable version of libxml2 (before 2.13.8) that contains a heap-based buffer under-read (CVE-2025-32415) in the xmlSchemaIDCFillNodeTables function in xmlschemas.c. The issue can be triggered when validating against an untrusted XML Schema, or when validating untrusted documents against trusted schemas that use xsd:keyref in combination with recursively defined types that have additional identity constraints. Upstream and MITRE rate this issue as low severity.

🎖@cveNotify

🚨 CVE-2025-71407
Nokogiri before 1.18.3 contains a stack buffer overflow vulnerability in libxml2 when reporting DTD validation errors with long QName prefixes, and a use-after-free vulnerability during validation against untrusted XML Schemas. Attackers can trigger these vulnerabilities by providing malicious DTD content or untrusted XSD files to cause denial of service or potential code execution.

🎖@cveNotify

🚨 CVE-2026-55532
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, MCP HTTP Stream _validate_origin uses request_origin.startswith(allowed), allowing the attacker-controlled localhost.attacker.com HTTP origin to satisfy the localhost allowlist. A webpage can send Content-Type: text/plain requests without preflight and invoke tools/call without an API key, including file writes that persist agent instructions. This issue is fixed in version 4.6.58.

🎖@cveNotify

🚨 CVE-2026-55533
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, create_auth_middleware() allows requests when auth=api-key lacks PRAISONAI_API_KEY or JWT authentication lacks PRAISONAI_JWT_SECRET. An externally bound Recipe server can therefore accept unauthenticated POST /v1/recipes/run requests despite authentication being enabled. This issue is fixed in version 4.6.58.

🎖@cveNotify