AI Security
LLM09 misinformation: what should you log and alert on?
Log the prompt, sources, model version and review outcome for consequential answers. Alert on ungrounded claims, unknown packages and unreviewed actions.
Oct 18, 20264 min read
Resources
Supply Chain Security, in plain English.
Deep dives, practical guides, and incident analyses from engineers who build Safeguard. No fluff, no vendor FUD — just what you need to ship secure software.
Filter
Articles
RSS feedAI Security
OWASP LLM09 misinformation: which controls actually work?
LLM09 covers confident but false model output. Grounding, validation and trained review help. Disclaimers alone do not. Start with your high-stakes uses.
Oct 18, 20264 min read
AI Security
OWASP LLM09 misinformation: a checklist for engineering leads
LLM09 covers confident but false model output. Ground answers in trusted sources, validate automatically, keep humans on high-stakes calls. A checklist.
Oct 18, 20265 min read
AI Security
How do you test a RAG app for OWASP LLM08 vector weaknesses?
Test LLM08 by checking who can retrieve what, what enters the knowledge base, and what gets logged. A test plan for RAG access, poisoning and leakage.
Oct 18, 20264 min read
AI Security
What should you log and alert on for LLM08 vector weaknesses?
Log who retrieved which chunks, where each came from and what group it belongs to. Alert on cross-tenant hits, new sources and unusual retrieval spikes.
Oct 18, 20264 min read
AI Security
OWASP LLM08 vector and embedding weaknesses: a checklist for leads
LLM08 covers how vectors and embeddings are generated, stored and retrieved in RAG systems. Review access, sources, tenancy, logging and behavior.
Oct 18, 20265 min read
AI Security
System prompt leakage (OWASP LLM07): what works and what doesn't
Treat the system prompt as public. Keep secrets and access rules out of it and enforce permissions outside the model. See which controls work and fail.
Oct 18, 20264 min read
AI Security
OWASP LLM07 system prompt leakage: a checklist for engineering leads
OWASP says treat the system prompt as readable. Use this checklist to move secrets out, enforce rules outside the model, and test for what leaks.
Oct 18, 20264 min read
AI Security
OWASP LLM06 excessive agency, explained with an email assistant
Excessive agency means an LLM agent can do more than it should. Learn the three root causes, OWASP's email assistant scenario and the first fixes.
Oct 18, 20265 min read
Stay informed
Weekly insights on software supply chain security, delivered to your inbox.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.