GHSA-Q399-23R3-HFX4 OpenClaw: system.run approvals did not bind PATH-token executable identity, enabling post-approval executable rebind

Summary For host=node runs, approvals validated command context but did not pin executable identity for non-path-like argv0 tokens for example tr. If PATH resolution changed after approval, execution could run a different binary. Impact A previously approved action could execute a different...

Server-side Request Forgery (SSRF)

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Server-side Request Forgery SSRF in the websearch citation redirect. An attacker can access internal network resources by supplying a crafted citation redirect target that points to...

GHSA-G99V-8HWM-G76G OpenClaw has web_search citation redirect SSRF via private-network-allowing policy

Summary Gemini websearch citation redirect resolution used a private-network-allowing SSRF policy. A citation URL redirect could target loopback/private/internal destinations and be fetched by the gateway. Impact An attacker who can influence citation redirect targets could trigger internal-netwo...

OpenClaw has web_search citation redirect SSRF via private-network-allowing policy

Summary Gemini websearch citation redirect resolution used a private-network-allowing SSRF policy. A citation URL redirect could target loopback/private/internal destinations and be fetched by the gateway. Impact An attacker who can influence citation redirect targets could trigger internal-netwo...

Missing Authorization

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Missing Authorization through an authorization mismatch in the agent. An attacker can perform privileged control-plane actions beyond their intended write scope by invoking owner-only too...

GHSA-7XMQ-G46G-F8PV OpenClaw: Sandbox media TOCTOU could read files outside sandbox root

Summary Sandbox media handling had a time-of-check/time-of-use gap: media paths could be validated first and read later through a separate path. A symlink retarget between those steps could cause reads outside sandboxRoot. Impact Affected versions could permit host file reads outside the intended...

Time-of-check Time-of-use (TOCTOU) Race Condition

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Time-of-check Time-of-use TOCTOU Race Condition via the media attachment handling. An attacker can access files outside the intended sandbox boundary by exploiting a race condition betwee...

Time-of-check Time-of-use (TOCTOU) Race Condition

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Time-of-check Time-of-use TOCTOU Race Condition via the writeFileWithinRoot function. An attacker can create or truncate files outside the intended root directory by exploiting a race...

Command Injection

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Command Injection in the wrapper resolution. An attacker can execute arbitrary commands by influencing the current working directory during wrapper resolution for .cmd or .bat files on...

GHSA-8M9V-XPGF-G99M OpenClaw has an unauthorized sender bypass in its stop triggers and /models command authorization

Summary Unauthorized senders could trigger two command paths without sender authorization checks: 1. stop-like natural-language abort triggers 2. /models command output Impact An unauthorized sender could disrupt active sessions and view model/auth metadata that should be authorization-gated. Fix...

Incorrect Authorization

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Incorrect Authorization in the stop triggers and /models command. An attacker can disrupt active sessions and access sensitive model or authentication metadata by sending unauthorized...

OpenClaw has an unauthorized sender bypass in its stop triggers and /models command authorization

Summary Unauthorized senders could trigger two command paths without sender authorization checks: 1. stop-like natural-language abort triggers 2. /models command output Impact An unauthorized sender could disrupt active sessions and view model/auth metadata that should be authorization-gated. Fix...

Access Control Bypass

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Access Control Bypass in the sessionsspawn sandboxed session. An attacker can bypass intended sandbox restrictions by spawning a child process under an agent with sandboxing disabled,...

OpenClaw's sandboxed sessions_spawn now enforces sandbox inheritance for cross-agent spawns

Summary A sandboxed session could use cross-agent sessionsspawn to create a child under an agent configured with sandbox.mode="off", downgrading runtime confinement. Impact In mixed-agent setups that allow cross-agent spawning, a sandboxed requester could escape into an unsandboxed child runtime...

GHSA-WR6M-JG37-68XH OpenClaw has unbounded memory growth in Zalo webhook via query-string key churn (unauthenticated DoS)

Summary Unauthenticated requests to a reachable Zalo webhook endpoint could trigger unbounded in-memory key growth by varying query strings on the same valid webhook route. Impact An attacker could cause memory pressure and potential process instability or OOM, degrading availability. Fix Webhook...

Allocation of Resources Without Limits or Throttling

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Allocation of Resources Without Limits or Throttling in the Zalo webhook. An attacker can exhaust system memory and cause process instability or termination by sending unauthenticated...

Interpretation Conflict

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Interpretation Conflict via the platform or deviceFamily metadata fields. An attacker can expand node command availability beyond intended defaults by supplying Unicode-confusable values...

GHSA-VPJ2-69HF-RPPW OpenClaw: Browser control startup could continue unauthenticated after auth bootstrap failure

Summary When browser control started without explicit auth credentials, OpenClaw attempted to bootstrap auth automatically. In affected versions, if that bootstrap step threw an error, startup could continue and expose browser-control routes without authentication. Impact On affected deployments,...

OpenClaw: Browser control startup could continue unauthenticated after auth bootstrap failure

Summary When browser control started without explicit auth credentials, OpenClaw attempted to bootstrap auth automatically. In affected versions, if that bootstrap step threw an error, startup could continue and expose browser-control routes without authentication. Impact On affected deployments,...

A week in security (February 23 – March 1)

Last week on Malwarebytes Labs: Public Google API keys can be used to expose Gemini AI data Inside a fake Google security check that becomes a browser RAT Fake Zoom and Google Meet scams install Teramind: A technical deep dive How to understand and avoid Advanced Persistent Threats The Conduent...