Writeup Exploits
63,779 exploits tracked across all sources.
OpenClaw 2026.4.5 < 2026.4.10 - Sandbox Escape via host Parameter Override in Exec Routing
OpenClaw versions 2026.4.5 before 2026.4.10 contain a sandbox escape vulnerability allowing sandboxed agents to override exec routing by specifying host=node. Attackers can bypass sandbox boundaries and route execution to remote nodes instead of intended sandbox paths.
CVSS 8.8
OpenClaw 2026.2.22 < 2026.4.12 - Shell-Wrapper Detection Bypass via Environment Variable Assignment Injection
OpenClaw versions from 2026.2.22 before 2026.4.12 contain an insufficient shell-wrapper detection vulnerability allowing attackers to inject environment variable assignments at the argv level. Attackers can bypass exec preflight handling to manipulate high-risk shell variables like SHELLOPTS and PS4, affecting execution semantics and security controls.
CVSS 8.8
OpenClaw < 2026.4.14 - Internal Page Content Exposure via Browser Snapshot and Screenshot Routes
OpenClaw before 2026.4.14 contains an improper access control vulnerability in browser snapshot, screenshot, and tab routes that fail to consistently validate the final browser target after navigation. Authenticated callers can bypass SSRF restrictions to expose internal or disallowed page content by exploiting route-driven navigation without proper policy re-validation.
CVSS 7.7
OpenClaw 2026.4.9 < 2026.4.10 - Denial of Service via Oversized WebSocket Frames in Voice-call Realtime Path
OpenClaw versions 2026.4.9 before 2026.4.10 contain a denial of service vulnerability in the voice-call realtime WebSocket path that accepts oversized frames without proper validation. Remote attackers can send oversized WebSocket frames to cause service unavailability for deployments exposing the voice-call realtime WebSocket path.
CVSS 7.5
OpenClaw 2026.4.9 < 2026.4.10 - Sender Policy Bypass in Host Media Attachment Reads
OpenClaw versions 2026.4.9 before 2026.4.10 contain a sender policy bypass vulnerability in the outbound host-media attachment read helper that allows unauthorized local file disclosure. Attackers with denied read access via toolsBySender or group policy can trigger host-media attachment loading to bypass sender and group-scoped authorization boundaries and retrieve readable local files through the outbound media path.
CVSS 7.7
OpenClaw < 2026.4.10 - SSRF Policy Bypass in Browser Tabs Action Routes
OpenClaw before 2026.4.10 contains a server-side request forgery policy bypass vulnerability in the browser tabs action select and close routes. Attackers can bypass configured browser SSRF policy protections by exploiting the /tabs/action endpoint to perform unauthorized tab navigation operations.
CVSS 8.5
OpenClaw < 2026.4.12 - Server-Side Request Forgery via QQBot Reply Media URL Handling
OpenClaw before 2026.4.12 contains a server-side request forgery vulnerability in QQBot reply media URL handling that allows attackers to fetch arbitrary content. Attackers can exploit this by providing malicious media URLs that trigger SSRF requests, with fetched bytes subsequently re-uploaded through the channel.
CVSS 8.2
OpenClaw < 2026.4.14 - Server-Side Request Forgery via Private Network Navigation
OpenClaw before 2026.4.14 contains a server-side request forgery vulnerability in browser SSRF policy that allows private-network navigation by default. Attackers can exploit this misconfiguration to access internal services or metadata endpoints through browser-driven requests.
CVSS 7.7
OpenClaw < 2026.4.14 - Redaction Bypass via sourceConfig and runtimeConfig Aliases
OpenClaw before 2026.4.14 contains a redaction bypass vulnerability that allows authenticated gateway clients to receive unredacted secrets through sourceConfig and runtimeConfig alias fields. Attackers with config read access can exploit this to obtain provider API keys, gateway authentication material, and channel credentials that should have been redacted.
CVSS 6.5
OpenClaw < 2026.4.10 - Time-of-Check-Time-of-Use (TOCTOU) Race Condition in exec Script Preflight Validator
OpenClaw before 2026.4.10 contains a time-of-check-time-of-use vulnerability in the validateScriptFileForShellBleed function that allows local attackers to bypass workspace boundary checks. An attacker with workspace write access can race-condition swap the target file between validation and preflight read, causing the validator to inspect a different file identity than the one that passed the initial boundary check.
CVSS 2.5
OpenClaw 2026.2.23 < 2026.4.12 - Weakened Exec Approval Binding via busybox and toybox Applet Execution
OpenClaw versions 2026.2.23 before 2026.4.12 contain a weakened exec approval binding vulnerability in busybox and toybox applet execution that allows attackers to obscure which applet would actually run. Attackers can exploit opaque multi-call binaries to bypass exec approval mechanisms and weaken risk classification of unsafe applet invocations.
CVSS 8.8
OpenClaw < 2026.4.9 - Environment Variable Injection via Workspace .env File
OpenClaw before 2026.4.9 contains an environment variable injection vulnerability allowing malicious workspace .env files to set runtime-control variables. Attackers can inject variables affecting update sources, gateway URLs, ClawHub resolution, and browser executable paths to compromise application behavior.
CVSS 7.3
OpenClaw 2026.4.7 < 2026.4.10 - Sandbox Media Normalization Bypass via Discord Event Cover Image
OpenClaw versions 2026.4.7 before 2026.4.10 fail to normalize Discord event cover image parameters in sandbox media processing. Attackers can bypass media normalization to inject host-local media references into channel action paths expecting normalized media.
CVSS 7.7
OpenClaw < 2026.4.10 - Arbitrary Local File Read via QQBot Media Tags
OpenClaw before 2026.4.10 contains an arbitrary file read vulnerability in QQBot media tags that allows attackers to reference host-local paths outside the intended media storage boundary. Attackers can craft malicious reply text containing media tags to disclose arbitrary local files through outbound media handling.
CVSS 8.6
OpenClaw < 2026.4.10 - Unsanitized External Input in Agent Hook Events
OpenClaw before 2026.4.10 contains an input validation vulnerability that allows external hook metadata to be enqueued as trusted system events. Attackers can supply malicious hook names to escalate untrusted input into higher-trust agent context.
CVSS 9.1
OpenClaw < 2026.4.14 - Authorization Context Reuse in Collect-Mode Queue Batches
OpenClaw before 2026.4.14 contains an authorization context reuse vulnerability in collect-mode queue batches that allows messages from different senders to inherit the final sender's authorization context. Attackers can exploit this by sending multiple queued messages to drain batches using a more privileged sender's context, causing earlier messages to execute with elevated permissions.
CVSS 6.8
OpenClaw 2026.4.7 < 2026.4.14 - Privilege Escalation via Untrusted Webhook Wake Events
OpenClaw versions 2026.4.7 before 2026.4.14 contain a privilege escalation vulnerability where heartbeat owner downgrade logic skips webhook wake events carrying untrusted content. Attackers can exploit this by sending untrusted webhook wake events to preserve owner-like execution context when the run should have been downgraded.
CVSS 9.1
OpenClaw < 2026.4.10 - Path Traversal in screen_record outPath Parameter
OpenClaw before 2026.4.10 contains a path traversal vulnerability in the screen_record tool's outPath parameter that bypasses workspace-only filesystem guards. Attackers can exploit this by specifying an outPath outside the workspace boundary to write files to unintended locations on the system.
CVSS 6.5
OpenClaw 2026.4.5 < 2026.4.10 - Privilege Escalation via Memory Dreaming Configuration in /dreaming Endpoint
OpenClaw versions 2026.4.5 before 2026.4.10 contain a privilege escalation vulnerability allowing write-scoped operators to modify persistent memory dreaming settings. Attackers with write-scoped gateway access can toggle admin-class configuration mutations through the /dreaming endpoint to escalate privileges.
CVSS 6.5
OpenClaw < 2026.4.9 - Untrusted Provider Plugin Auto-enablement via Workspace Provider Auth
OpenClaw before 2026.4.9 contains an authentication bypass vulnerability allowing untrusted workspace plugins to be auto-enabled during non-interactive onboarding when provider auth choices are shadowed. Attackers can exploit this by crafting malicious workspace plugins that are automatically selected and enabled during authentication setup without explicit user consent.
CVSS 8.8
OpenClaw 2026.3.22 < 2026.4.5 - Symlink Traversal in Remote Marketplace Repository Path Handling
OpenClaw versions 2026.3.22 before 2026.4.5 contain a symlink traversal vulnerability in remote marketplace repository path handling that allows attackers to escape the expected repository root. Attackers can exploit this by providing crafted symlink paths to access files outside the intended repository directory.
CVSS 6.5
OpenClaw < 2026.4.10 - Untrusted Workspace Plugin Shadow Resolution in Channel Setup
OpenClaw before 2026.4.10 contains a plugin trust bypass vulnerability that allows channel setup catalog lookups to resolve workspace plugin shadows before bundled channel plugins. Attackers can exploit this by crafting malicious workspace plugins that bypass intended trust gates during setup-time plugin loading.
CVSS 8.8
OpenClaw 2026.4.10 < 2026.4.14 - Missing Sender Authorization in Microsoft Teams SSO Invoke Handler
OpenClaw versions 2026.4.10 before 2026.4.14 contain a missing authorization vulnerability in the Microsoft Teams SSO invoke handler that fails to apply sender allowlist checks. Attackers can bypass sender authorization by sending SSO invoke requests that are processed without proper validation, allowing unauthorized access to Teams SSO signin functionality.
CVSS 5.3
OpenClaw < 2026.4.10 - SSRF Policy Bypass in Existing-Session Browser Interaction Routes
OpenClaw before 2026.4.10 contains a server-side request forgery policy bypass vulnerability in existing-session browser interaction routes. Attackers can bypass SSRF navigation guards to interact with or navigate to unauthorized targets without policy enforcement.
CVSS 7.7
OpenClaw < 2026.4.12 - Improper Authorization via Empty Approver Lists
OpenClaw before 2026.4.12 contains an improper authorization vulnerability in helper-backed channels where empty resolved approver lists are interpreted as explicit approval authorization. Attackers can resolve pending approvals without proper authorization by exploiting this logic flaw if they know an approval id.
CVSS 6.5
By Source