Advisories
Flaws in things people run, newest first. The level comes from evidence, never from how alarming it sounds.
- This weekExploitable · Confirmedpillow: Pillow EpsImagePlugin negative %%BeginBinary byte count causes infinite loop denial of service
Fix: pillow 12.3.0
CVE-2026-59203 ·
- This weekExploitable · Confirmedpraisonaiagents: PraisonAI: Human-in-the-loop tool approval is cached by tool name and silently reused for all subsequent calls with arbitr…
Fix: praisonaiagents 1.6.78
CVE-2026-60087 ·
- This weekExploitable · Confirmed@vitest/browser: Vitest browser mode serves unsanitized otelCarrier query parameter as inline script
Fix: @vitest/browser 5.0.0-beta.3 / 4.1.6
CVE-2026-47428 ·
- This weekExploitable · Confirmeddompurify: DOMPurify: Cross-realm IN_PLACE sanitization leaves executable markup intact via realm-bound `instanceof` checks
Fix: dompurify 3.4.6
CVE-2026-49458 ·
- This weekExploitable · Confirmedsigstore: sigstore's `certificateOIDs` verification constraints are silently dropped and never enforced
Fix: sigstore 4.1.1
CVE-2026-48815 ·
- This weekExploitable · Confirmedlinkify-it: LinkifyIt#match scan loop has quadratic algorithmic complexity
Fix: linkify-it 5.0.1
CVE-2026-48801 ·
- This weekExploitable · Confirmedua-parser-js: UAParser.js: Unbounded `Sec-CH-UA-Model` parsing can trigger ReDoS in `withClientHints()`
Fix: ua-parser-js 2.0.10
CVE-2026-48125 ·
- This weekExploitable · Confirmed@sigstore/core: @sigstore/core has DSSE payloadType type-binding failure
Fix: @sigstore/core 3.2.1
CVE-2026-48758 ·
- This weekExploitable · Confirmedtwig/twig: Twig: The `spaceless` filter implicitly marks its output as safe
Fix: twig/twig 3.26.0
CVE-2026-46628 ·
- This weekExploitable · Confirmedgithub.com/rclone/rclone: rclone: Unvalidated symlink target in local `--links` — arbitrary file write from an untrusted remote
Fix: github.com/rclone/rclone 1.74.4
CVE-2026-54572 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: AgentMail webhook mode accepts forged unsigned message.received events and invokes agents
Fix: praisonai 4.6.78
CVE-2026-61436 ·
- This weekExploitable · Confirmedpraisonaiagents: PraisonAI: DNS rebinding bypass in `web_crawl` SSRF protection allows internal response disclosure
Fix: praisonaiagents 1.6.78
CVE-2026-61430 ·
- This weekExploitable · Confirmedpraisonaiagents: PraisonAI: SkillTools Executes Scripts Without Path Containment Validation
Fix: praisonaiagents 1.6.78
CVE-2026-61443 ·
- This weekExploitable · Confirmedpillow: Pillow TGA RLE encoder can serialize up to ~57 KB of adjacent heap data into generated images
Fix: pillow 12.3.0
CVE-2026-59198 ·
- This weekExploitable · Confirmedpillow: Pillow: Decompression Bomb DoS via PdfParser.PdfStream.decode()
Fix: pillow 12.3.0
CVE-2026-59200 ·
- This weekExploitable · Confirmedpillow: Pillow: Out-of-bounds read via attacker-controlled row stride on Pillow's mmap path (McIdas AREA files)
Fix: pillow 12.3.0
CVE-2026-54058 ·
- This weekExploitable · Confirmedpillow: Pillow: Controlled heap out-of-bounds write in Pillow `ImageCmsTransform.apply()` via output mode mismatch
Fix: pillow 12.3.0
CVE-2026-59205 ·
- This weekExploitable · Confirmedalextselegidis/easyappointments: Easy!Appointments disable_booking_message rendered as raw HTML on public booking page — Stored XSS
Fix: alextselegidis/easyappointments 1.6.1
CVE-2026-52838 ·
- This weekExploitable · Confirmedalextselegidis/easyappointments: Easy!Appointments Vulnerable to Appointments Takeover via Excessive Data Exposure
No fixed release is listed yet.
CVE-2026-55651 ·
- This weekExploitable · Confirmedpyod: PyOD persistence.load deserializes untrusted artifacts before validation
Fix: pyod 3.6.2
CVE-2026-15529 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: PGVector and Cassandra knowledge stores interpolate vector dimensions into DDL
Fix: praisonai 4.6.78
CVE-2026-60090 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: AICoder Arbitrary File Write and Command Execution via LLM Tool Calls
Fix: praisonai 4.6.78
CVE-2026-61445 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: Prompt Injection Defense Bypassed for HIGH-Level Threats
Fix: praisonai 4.6.78
CVE-2026-61439 ·
- This weekExploitable · Confirmedappium-mcp: appium-mcp: Unescaped Locator Data XSS in MCP-UI Resource (createLocatorGeneratorUI)
Fix: appium-mcp 1.85.10
CVE-2026-58500 ·
- This weekExploitable · Confirmedsnipe/snipe-it: Snipe-IT has an authorization bypass on bulk editing users
Fix: snipe/snipe-it 8.6.2
CVE-2026-55460 ·
- This weekExploitable · Confirmedcrewai-tools: crewai-tools SSRF redirect bypass exposes internal services
Fix: crewai-tools 1.15.1
CVE-2026-62240 ·
- This weekExploitable · Confirmedkeras: Keras: tar extraction permits symlink-based path traversal
Fix: keras 3.15.0 / 3.12.3
CVE-2026-12482 ·
- This weekExploitable · Confirmedgetgrav/grav: Grav: Admin Backup Zip File Exposes Account Credentials and Configuration Secrets
Fix: getgrav/grav 1.7.53
CVE-2026-55885 ·
- This weekExploitable · Confirmedgithub.com/xuri/excelize/v2, github.com/xuri/excelize: Excelize: Streaming GetRows row-bound bypass causes attacker-controlled allocation
Fix: github.com/xuri/excelize/v2 2.11.0
CVE-2026-59161 ·
- This weekExploitable · Confirmedpraisonaiagents: PraisonAI: Unsafe Dynamic Module Loading Leads to Arbitrary Code Execution via tools.py in AgentFlow
Fix: praisonaiagents 1.6.78
CVE-2026-61437 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: Prompt-injection defense blocks only when 3+ detector families fire simultaneously; realistic single-vector injection…
Fix: praisonai 4.6.78
CVE-2026-60086 ·
- This weekExploitable · Confirmedlangroid: Langroid: Path traversal in the file tools allows read/write outside configured current directory
Fix: langroid 0.64.0
CVE-2026-50181 ·
- This weekExploitable · Confirmedopen-webui: Open WebUI: Upload `metadata.knowledge_id` bypasses the knowledge-base write-access check (read-only users can add files to KB)
Fix: open-webui 0.10.0
CVE-2026-59217 ·
- This weekExploitable · Confirmedopen-webui: Open WebUI: POST /api/v1/images/edit bypasses the global image-edit switch and the per-user image-generation permission
Fix: open-webui 0.10.0
CVE-2026-59227 ·
- This weekExploitable · Confirmedsnipe/snipe-it: Snipe-IT vulnerable to stored XSS via inline-served attachment
Fix: snipe/snipe-it 8.6.2
CVE-2026-55466 ·
- This weekExploitable · Confirmed9router: 9router: Image prefetch DNS rebinding allows SSRF to internal services
Fix: 9router 0.5.2
CVE-2026-56676 ·
- This weekExploitable · Confirmedphalcon/cphalcon: Phalcon: Catastrophic backtracking (ReDoS) in the default Phalcon Router route lead to remote unauthenticated DoS
Fix: phalcon/cphalcon 5.15.0
CVE-2026-57584 ·
- This weekExploitable · Confirmedprowler-cloud: Prowler: SAML Domain Claiming Enables Cross-Tenant Account Takeover
Fix: prowler-cloud 5.30.3
CVE-2026-59151 ·
- This weekExploitable · Confirmedgithub.com/nezhahq/nezha: Nezha Dashboard: DDNS and Notification credential exposure via unredacted list API
Fix: github.com/nezhahq/nezha 2.2.5
CVE-2026-59155 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: Project custom command templates can read outside-workspace files into model prompts
Fix: praisonai 4.6.78
CVE-2026-60088 ·
- This weekExploitable · Confirmedpraisonai-platform: PraisonAI: Platform member PATCH routes allow owner resource rewrites and project lead reassignment delete bypass
Fix: praisonai-platform 0.1.9
CVE-2026-61442 ·
- This weekExploitable · Confirmedgithub.com/gotenberg/gotenberg/v8, github.com/gotenberg/gotenberg/v7: Gotenberg: SSRF via LibreOffice document processing
Fix: github.com/gotenberg/gotenberg/v8 8.34.0
CVE-2026-55229 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: AgentMail webhook lacks signature verification, allowing unauthenticated message injection and sender spoofing
Fix: praisonai 4.6.78
CVE-2026-61428 ·
- This weekExploitable · Confirmedgithub.com/axllent/mailpit: Mailpit: Incomplete SSRF protection in Link Check API via IPv6 transition mechanisms
Fix: github.com/axllent/mailpit 1.30.2
CVE-2026-55187 ·
- This weekExploitable · Confirmedpraisonaiagents: PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets
Fix: praisonaiagents 1.6.78
CVE-2026-61447 ·
- This weekExploitable · Confirmedpraisonai: PraisonAI: AgentOS defaults to network-exposed no-auth mode, allowing unauthenticated agent invocation and instruction disclosure
Fix: praisonai 1.7.3
CVE-2026-61426 ·
- This weekExploitable · Confirmedsnipe/snipe-it: Snipe-IT: Cross-company deletion of pending checkout acceptances via unscoped report endpoint
Fix: snipe/snipe-it 8.6.2
CVE-2026-55515 ·
- This weekExploitable · Confirmedsnipe/snipe-it: Snipe-IT has an authorization bypass on print inventory page
Fix: snipe/snipe-it 8.6.1
CVE-2026-55462 ·
- This weekExploitable · Confirmedmcp-server-kubernetes: mcp-server-kubernetes argument injection can expose Kubernetes cluster credentials
Fix: mcp-server-kubernetes 3.9.0
CVE-2026-61459 ·
- This weekExploitable · Confirmedpraisonaiagents: PraisonAI: Crawl4AI/Chromium backend is also affected by the `web_crawl` SSRF validation bypass
Fix: praisonaiagents 1.6.78
CVE-2026-61429 ·
Follow without an account:everything followedact nowthis weekJSON