Summary When the Slack integration is enabled, Slack channel metadata (topic/description) could be incorporated into the model's system prompt.

Impact Prompt injection is a documented risk for LLM-driven systems. This issue increased the injection surface by allowing untrusted Slack channel metadata to be treated as higher-trust system input.

This is relevant only for deployments that enable Slack. In deployments where tool execution is enabled, a successful injection could lead to unintended tool invocations and/or unintended data exposure.

Affected Packages / Versions - npm: openclaw < 2026.2.3

Patched Versions - npm: openclaw >= 2026.2.3

Mitigation - If you do not use Slack: no action required. - If you use Slack: upgrade to a patched version.

Fix Commit(s) - 35eb40a7000b59085e9c638a80fd03917c7a095e

Thanks @KonstantinMirin for reporting.

A weakness has been identified in OpenClaw up to 2026.1.26. Affected by this issue is some unknown functionality of the file src/agents/tools/web-fetch.ts of the component assertPublicHostname Handler. Executing a manipulation can lead to server-side request forgery. The attack can be executed remotely. This attack is characterized by high complexity. The exploitation is known to be difficult. The exploit has been made available to the public and could be used for attacks. Upgrading to version 2026.1.29 can resolve this issue. This patch is called b623557a2ec7e271bda003eb3ac33fbb2e218505. Upgrading the affected component is advised.

OpenClaw versions prior to 2026.2.23 contain an html injection vulnerability in the HTML session exporter that allows attackers to execute arbitrary javascript by injecting malicious mimeType values in image content blocks. Attackers can craft session entries with specially crafted mimeType attributes that break out of the img src data-URL context to achieve cross-site scripting when exported HTML is opened.

OpenClaw versions prior to 2026.2.21 accept prototype-reserved keys in runtime /debug set override object values, allowing prototype pollution attacks. Authorized /debug set callers can inject proto, constructor, or prototype keys to manipulate object prototypes and bypass command gate restrictions.

OpenClaw versions prior to 2026.2.26 contain an authorization bypass vulnerability where DM pairing-store identities are incorrectly treated as group allowlist identities when dmPolicy=pairing and groupPolicy=allowlist. Remote attackers can send messages and reactions as DM-paired identities without explicit groupAllowFrom membership to bypass group sender authorization checks.

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 webfetch functionality to access blocked addresses such as 198.18.0.0/15 and other non-global ranges.

OpenClaw versions prior to 2026.2.22 fail to consistently validate redirect chains against configured mediaAllowHosts allowlists during MSTeams media downloads. Attackers can supply or influence attachment URLs to force redirects to non-allowlisted targets, bypassing SSRF boundary controls.

OpenClaw before 2026.3.8 contains a sender allowlist bypass vulnerability in its Microsoft Teams plugin that allows unauthorized senders to bypass intended authorization checks. When a team/channel route allowlist is configured with an empty groupAllowFrom parameter, the message handler synthesizes wildcard sender authorization, permitting any sender in the matched team/channel to trigger replies in allowlisted Teams routes.

OpenClaw before 2026.3.25 contains an authorization bypass vulnerability in Google Chat group policy enforcement that relies on mutable space display names. Attackers can rebind group policies by changing or colliding space display names to gain unauthorized access to protected resources.

OpenClaw before 2026.3.22 contains a policy confusion vulnerability in room authorization that matches colliding room names instead of stable room tokens. Attackers can exploit similarly named rooms to bypass allowlist policies and gain unauthorized access to protected Nextcloud Talk rooms.

OpenClaw before 2026.3.22 contains a policy bypass vulnerability where queued node actions are not revalidated against current command policy when delivered. Attackers can exploit stale allowlists or declarations that survive policy tightening to execute unauthorized commands.

Affected Packages / Versions

- Package: openclaw (npm) - Affected versions: < 2026.4.20 - Patched version: 2026.4.20

Impact

The Control UI assistant-media route authenticated trusted-proxy callers but did not enforce the declared operator scopes for identity-bearing HTTP auth paths. A trusted-proxy caller without operator.read could access assistant-media files and metadata that were otherwise inside allowed media roots.

The route still required successful gateway authentication and media-root checks. Severity is low.

Fix

Assistant-media file and metadata requests now require operator.read on identity-bearing HTTP auth paths.

Fix commit:

- 99ef3a63c58440d53f8e45ad861b846032fcb036

Release

Fixed in OpenClaw 2026.4.20.

OpenClaw before 2026.3.31 lacks browser-origin validation in HTTP operator endpoints when operating in trusted-proxy mode, allowing cross-site request forgery attacks. Attackers can exploit this by sending malicious requests from a browser in trusted-proxy deployments to perform unauthorized actions on HTTP operator endpoints.

OpenClaw before 2026.3.31 contains a logic error in Discord component interaction routing that misclassifies group direct messages as direct messages in extensions/discord/src/monitor/agent-components-helpers.ts. Attackers can exploit this misclassification to bypass group DM policy enforcement or trigger incorrect session handling.

OpenClaw before 2026.3.31 contains an authorization bypass vulnerability in Discord slash command and autocomplete paths that fail to enforce group DM channel allowlist restrictions. Authorized Discord users can bypass channel restrictions by invoking slash commands, allowing access to restricted group DM channels.

Summary

Before OpenClaw 2026.4.2, Slack thread starter and thread-history context fetched through the API was not filtered by the effective sender allowlist. Messages from non-allowlisted senders could still enter the agent context when an allowlisted user replied in the same thread.

Impact

A Slack deployment that relied on sender allowlists could still feed non-allowlisted thread content into the model context through thread history. This was a sender-access-control bypass on Slack thread context, not a direct channel-auth bypass.

Affected Packages / Versions

- Package: openclaw (npm) - Affected versions: <= 2026.4.1 - Patched versions: >= 2026.4.2 - Latest published npm version: 2026.4.1

Fix Commit(s)

- ac5bc4fb37becc64a2ec314864cca1565e921f2d — filter Slack thread context by the effective allowlist

Release Process Note

The fix is present on main and is staged for OpenClaw 2026.4.2. Publish this advisory after the 2026.4.2 npm release is live.

OpenClaw thanks @AntAISecurityLab for reporting.

OpenClaw before 2026.3.31 fails to terminate active WebSocket sessions when rotating device tokens. Attackers with previously compromised credentials can maintain unauthorized access through existing WebSocket connections after token rotation.

OpenClaw versions 2026.2.19 before 2026.3.31 contain an improper cache isolation vulnerability in the Zalo webhook replay-dedupe mechanism that is shared across authenticated webhook targets. Attackers controlling one authenticated Zalo webhook path in multi-account deployments can suppress legitimate events on different accounts by matching eventname and messageid parameters.

OpenClaw before 2026.3.31 contains an authorization bypass vulnerability in Discord voice ingress that allows attackers to bypass channel and member allowlist restrictions. Attackers can exploit stale-role validation gaps and improper channel name validation to gain unauthorized access to restricted voice channels.

OpenClaw before 2026.3.31 contains a scope bypass vulnerability in webhook replay cache deduplication that allows authenticated attackers to replay messages across sibling targets using the same messageId. Attackers can exploit overly broad cache keying to bypass replay protection and deliver duplicate webhook messages to unintended targets.

OpenClaw before 2026.3.31 contains a resource exhaustion vulnerability in media downloads that bypasses core safety limits for file size, count, and cleanup operations. Attackers can exhaust disk space by downloading media files without triggering intended safety restrictions, causing availability impact.

OpenClaw before 2026.4.8 omits owner-only enforcement for cross-channel allowlist writes in the /allowlist endpoint. An authorized non-owner sender can bypass access controls to perform allowlist modifications against different channels, violating the intended trust model.

OpenClaw before 2026.3.31 contains a sender allowlist bypass vulnerability that allows remote attackers to access restricted messages. Attackers can exploit fetched quoted, root, and thread context messages to bypass sender allowlist restrictions and retrieve unauthorized content.

OpenClaw before 2026.3.31 contains an access control bypass vulnerability in the Discord voice manager that allows attackers to bypass channel-level member access allowlist restrictions. Attackers can send Discord voice ingress requests before channel allowlist authorization is performed, gaining unauthorized access to restricted voice channels.

OpenClaw before 2026.3.31 contains an allowlist bypass vulnerability in Matrix thread root and reply context handling that fails to properly validate message senders. Attackers can fetch thread-root and reply context messages that should be filtered by sender allowlists, bypassing access controls.

OpenClaw before 2026.4.8 contains an authentication state management vulnerability where the resolvedAuth closure becomes stale after configuration reload. Newly accepted gateway connections continue using outdated resolved auth state, allowing attackers to bypass authentication controls through config reload operations.

OpenClaw before 2026.4.8 contains a session management vulnerability where existing WebSocket sessions survive shared gateway token rotation. Attackers can maintain unauthorized access to WebSocket connections after token rotation by exploiting the failure to disconnect existing shared-token sessions.

OpenClaw before 2026.4.15 contains an arbitrary file read vulnerability in the QMD backend memoryget function that allows callers to read any Markdown files within the workspace root. Attackers with access to the memory tool can bypass path restrictions by providing arbitrary workspace Markdown paths to read files outside canonical memory locations or indexed QMD result sets.

OpenClaw before 2026.4.22 contains a security envelope constraint bypass vulnerability allowing restricted subagents to spawn ACP child sessions that fail to inherit depth, child-count limits, control scope, or target-agent restrictions. Attackers can exploit this by spawning child sessions that bypass subagent-only constraints, potentially escalating privileges or accessing restricted resources.

OpenClaw before 2026.4.21 contains an authorization bypass vulnerability in command-auth.ts that allows non-owner senders to execute owner-enforced slash commands when wildcard inbound senders are configured without explicit owner allowFrom settings. Attackers can exploit this by sending commands like /send, /config, or /debug on affected channels to bypass owner-only command authorization checks.