Knowledge library
Articles
Engineering decisions, implementation notes, and operational field guides.
3 entries
AI pair programmingdeveloper telemetrypolicy enforcement
Telemetry-Driven Guardrails for AI Pair Programming
A practical architecture for instrumenting IDE assistants and coding agents to detect policy drift, evaluate engineering outcomes, and enforce security and governance controls without collecting source code or prompts.
Harish KumarAug 29, 2026
prompt injectionLLM securitycode assistants
Hardening LLM Code Assistants Against Prompt Injection
A practical threat model and defense-in-depth guide for reducing prompt injection risk in AI pair-programming workflows while preserving human control over code, tools, and production changes.
Harish KumarAug 28, 2026
prompt engineeringenterprise AIAI risk management
Why Prompt Engineering Is Not an Enterprise AI Skill Strategy
Prompt techniques can improve individual outputs, but enterprise AI adoption depends on verification, workflow redesign, critical thinking, risk controls, and domain judgment. Here is a practical capability model for engineering leaders.
Harish KumarAug 25, 2026