Topics
Browse posts by category and tag — every topic we cover, with the latest pieces under each.
Tags
- #mlops 5
- #observability 4
- #agents 3
- #agent-observability 2
- #governance 2
- #identity 2
- #blue-team 1
- #detection-engineering 1
- #formal-methods 1
- #incident-response 1
- #inference 1
- #langsmith 1
- #llm-monitoring 1
- #llm-security 1
- #local-llm 1
- #meta 1
- #mitre-atlas 1
- #monitoring 1
- #multi-agent 1
- #opentelemetry 1
- #phoenix 1
- #runbook 1
- #security 1
- #serving 1
- #siem 1
- #telemetry 1
- #tooling 1
- #tracing 1
Categories
monitoring 3 posts
- A Lean 4 stability proof for tool-mediated LLM agents, and what it means for your runbookA new arXiv paper certifies controllability and ISS robustness for an LLM-driven SOC agent using Lean 4. The MLOps takeaway is simpler than the math: monitor the action catalog, not the model.
- Embedding-Based Agent Monitoring Has a Blind Spot. Here's What to Watch Instead.A new paper demonstrates three attack patterns — Slow Drift, Benign Wrapper, Chaos Seeding — that defeat embedding-based detection of malicious agents in LLM multi-agent systems. The fix requires monitoring logit-level confidence, not just output embeddings.
- The Authority Gap Is an Observability Problem: What MLOps Teams Should BorrowA new framing of AI agent risk argues that delegation, not identity, is the missing telemetry. ML platform teams already have the substrate to fix it.
deep-dive 2 posts
- The Authority Gap Is an Observability Problem: What MLOps Teams Should Actually InstrumentOrchid Security's framing of agent governance as a delegation problem lands in the lap of ML observability teams. The instrumentation we already own decides whether the authority graph is real or theatre.
- The Agent Authority Gap Is an Observability Problem in a Security CostumeSecurity vendors are pitching 'continuous observability' as the answer to ungoverned AI agents. ML platform teams already shipped most of the pipes. The missing piece is identity context inside the trace span — and that is a schema fight, not a tooling fight.