Legion Command Papers
Agents are becoming infrastructure. The question is who governs them. These papers examine what changes structurally when AI moves from tool to operator in the world’s most demanding environments, and what it takes to deploy, govern, and prove those systems under real authority.
Blogs Posts
11 minute read
https://legionintel.com/command-papers/
why-retrieval-agents-fail
Why Retrieval Agents Fail: It’s Not Just the Model
Retrieval-focused agents fail for reasons beyond model capability. Performance is bounded by three independent constraints: the model, the retrieval stack, and prompt guidance. Evaluation finds the real bottleneck.
https://legionintel.com/team/
william-du
10 minute read
https://legionintel.com/command-papers/
ai-agent-evaluation-building-an-evaluation-platform-that-scales
AI Agent Evaluation: Building an Evaluation Platform That Scales
Teams deploy agents faster than they can test them. A single prompt change can silently degrade three agents while improving one. Here's how immutable datasets, purpose-driven metrics, and vendor-agnostic design make agentic evaluation reproducible at scale.
https://legionintel.com/team/
jennifer-sikos
12 minute read
https://legionintel.com/command-papers/
agentic-ai-security-architecture
Agentic AI Security is an Architectural Decision
Agentic AI security is an architectural property, not a feature you add after deployment. Governed workflows, strong typing, least-privilege access, and ephemeral agents address 20 OWASP vulnerabilities by design, validated at IL2 through IL6.
https://legionintel.com/team/
nick-weir
