Openclaw CVE-2026-41301

The exploit requires the target OpenClaw instance to have the Nostr DM ingress feature enabled and accessible from the attacker's network (default configuration for public Nostr relays). The vulnerability is triggered when the DM ingress receives any Nostr event with a valid Nostr event structure (including correctly formed event ID, kind field, and signature field) but with a forged or invalid signature that claims a victim's public key as the sender. No authentication to OpenClaw is required, and the attacker does not need to control any legitimate Nostr keys. The flaw is present only in the pairing challenge code path; DMs sent to already-paired peers are not affected (they use a separate, signature-validated ingress). Exploitation is limited to the pairing capacity of the target relay (shared resource pool); once pairing capacity is exhausted, further forgeries do not trigger additional work, resulting in a bounded DoS rather than infinite resource consumption. The vulnerability does NOT allow an attacker to impersonate an already-paired peer or decrypt legitimate paired DMs, as those features rely on cryptographic validation outside the vulnerable code path.

CVSS 6.9 (Network vector, low complexity, unauthenticated, no privilege requirement) indicates moderate network-exploitable risk with availability impact (VA:L reflects bounded denial of service). EPSS data not provided, but the attack vector (AV:N/AC:L/PR:N) combined with low technical complexity suggests high exploitability probability. No KEV listing or public POC reported, indicating this is not yet confirmed as actively exploited in the wild, though the straightforward attack (sending unauthenticated forged Nostr events) requires minimal sophistication. The real-world risk is bounded: while availability impact is confirmed (resource exhaustion via capacity flooding), the vulnerability does not enable code execution, authentication bypass for actual Nostr operations (only state creation), or confidentiality/integrity compromise of existing communications. Organizations running unpatched OpenClaw 2026.3.22-2026.3.30 on public Nostr relay networks face DoS risk proportional to their paired relay capacity; internal/firewalled deployments have lower risk. The AC:L and PR:N factors confirm this is not a theoretical vulnerability-any network path to the Nostr DM ingress permits exploitation.

An attacker on any network path to an OpenClaw instance generates and sends 1000 forged Nostr direct message events claiming to be from distinct victim Nostr public keys, each containing a pairing challenge request. OpenClaw's unpatched DM ingress processes these events before verifying the signature, creating 1000 pending pairing entries that consume the relay's shared pairing capacity. The relay then attempts to route pairing-reply messages back to the spoofed sender addresses and logs each transaction, causing unbounded logging work and relay bandwidth depletion. Legitimate peer pairing attempts are rejected once capacity is exhausted, resulting in denial of service for the pairing subsystem.