OpenClaw
Newest record 2026-10-27 · 2 distinct CVE ids · open in the explorer (the explorer matches the first product a record names, so it can list fewer)
Exploited or KEV-listed · 4
Records at severity 3–4 (exploitation confirmed by CISA, the vendor or reporting) and records flagged as listed in CISA KEV. The KEV flag is the classifier’s cisa-kev tag or a source on the KEV catalog; it is recorded when the collector sees the listing, so a missing flag does not mean a CVE is absent from the catalog.
- closedHermes Multi-Agent Framework Compromises Taiwanese Government Systems
Exploited · since 1 Jul 2026
AI - closedHermes Suspected Hackers Used AI Agents to Compromise Taiwanese Systems
Real loss · since 1 Jul 2026
AI - closedOpenClaw Exposed ClawdBot Control Interfaces Lead to Credential Access
Exploited · since 25 Jan 2026
AI - closedOpenClaw Malicious OpenClaw Skills Distribute AMOS Stealer via ClawHub
Exploited · since 1 Feb 2026
AI
Disclosures by month · last 24 months
Disclosed CVEs, no confirmed exploitation · 1
- closed8.6OpenClaw CVE-2026-100585: OpenClaw bypasses owner-only authorization for Claude Code prompts
Disclosed · since 26 Sep 2026 · CVE-2026-100585
AI
Other incidents: research and unexploited reports · 6
- closedOpenClaw Scope-Gap Injection: Exploiting Evaluator Blind Spots in OpenClaw
Research · since 27 Oct 2026
AI - closedOpenClaw Empirical Study of Security Concerns in OpenClaw Agentic Systems
Research · since 27 Oct 2026
AI - closedOpenClaw AutoRedClaw: Automated Red-Teaming for Persistent Claw-Style Agents
Research · since 21 Oct 2026
AI - closedOpenClaw Authorization-Grounded Workflow Attacks on Tool-Using Agents
Research · since 19 Oct 2026
AI - closed8.8OpenClaw OpenClaw Remote Code Execution Vulnerability Demonstrated
Research · since 1 Feb 2026 · CVE-2026-25253
AI - closedOpenClaw OpenClaw Command and Control via Indirect Prompt Injection
Research · since 3 Feb 2026
AI
Fix status and fixed versions are recorded when a source states them (each record’s page shows what is known); check the vendor’s own advisory before acting.