Markdown skill files propagate jailbreaks into AI agents
A poisoned SKILL.md raised mean attack success from 0.5% to 45% across seven open-weight models.
56
Product
LangChain
- Mentions · 7d0
- Mentions · 30d3
- All mentions7
- First seenAug 21, 2026
- Last mentioned13d ago
Timeline
7 articles · newest firstLarger gpt-oss agents gain no MCP fault tolerance, study finds
MCP-Injector, calibrated on 57,240 threads from 1,774 repositories, cut task scores 60 percent, from 8.08 to 3.23.
46
Jev as a judge: How counterfeit policy steers Jev
Datadog tests whether prompt injection can steer TypeSafe’s Jev and bypass LangChain tool-risk gating.
61
[AINews] Opus 5.5 is good at explainer videos
Claude Opus 5.5 leads SimpleBench at 88.4% as GPT-6, Gemini 3.8 Flash, and MiMo-V2.6-Pro also ship.
80
[AINews] Here are 6 Clones of Jev in 2 days
Jev's non-generative decision model launch drew 36M views and six open clones in two days, positioning discriminative models as fast complements to LLMs.
50
Arcjet brings security controls and audit trails to AI agents
Arcjet launched agent runtime security, giving engineering and security teams observability, policy enforcement, and audit trails for AI agents in production.
40
Wazuh and AI For Enhanced SOC Workflows
Wazuh details AI-powered SOC workflows via its AI Analyst, self-hosted Llama 3 via Ollama, and Claude 3.5 Haiku integrations.
25
Appears with
- Anthropicvendor · 3
- OpenAIvendor · 3
- Clineproduct · 2
- Metavendor · 2
- Jevai model · 2
- Claude Codeproduct · 2
- LangChainorganization · 2
- OpenAI Agents SDKproduct · 2
- Arcjetvendor · 1
- Vercelorganization · 1
- Gemini 3.8 Flashai model · 1
- Llama 3ai model · 1
- SemIfai model · 1
- Qwen3.8 Maxai model · 1
- OpenTelemetryproduct · 1
- Qwen3.5-9Bai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.