AgentArmor: A Framework, Evaluation, and Mitigation of Coding Agent Failures

arXiv·medium signal

Categorizes coding-agent failures into underspecification, capability errors, and harness errors, then proposes AgentArmor: an extended system prompt, a separate command classifier, a '3 strikes' policy, deterministic guardrails, and tools letting the agent edit its own context. Across 8 evaluations spanning 20 environments and 59 synthetic templates it shows statistically significant safety improvements. A practical playbook for hardening production coding agents.