OpenClaw AI Agent Flaws Could Enable Prompt Injection and Data Exfiltration
2026-03-15T13:24:12Z•42939b5c07b4d374d38b1ca4d443ba6b5574b830d90241408d40722bfa09727f
AI-agentchromecontainer-escapecredential-theftglasswormlaw-enforcement-takedownmalwaren8nnation-stateopenclawpatchingprompt-injectionremote-code-executionsupply-chainveeamvulnerabilityzero-day
What happened
A broad set of high-severity security incidents and patches were reported: CNCERT warned that OpenClaw (self‑hosted AI agent) has weak default security enabling prompt injection and data exfiltration; a GlassWorm supply‑chain campaign abused 72 Open VSX extensions to target developers; and a suspected China-linked espionage campaign is using AppleChris and MemFun against Southeast Asian militaries. Multiple actively exploited and critical vulnerabilities were disclosed and patched — including Chrome zero-days (Skia CVE-2026-3909), high‑severity n8n RCEs (CVE-2025-68613, CVE-2026-27577, CVE-202
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- the_hacker_news
- Record identifier
- 42939b5c07b4d374d38b1ca4d443ba6b5574b830d90241408d40722bfa09727f
- Enrichment time
- 2026-03-15T13:24:12Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.