Agent Loop: Reliability & Verification

An engineering series exploring execution safety, tool correctness, retrieval reliability, evidence verification, and independent testing in local-LLM agents.

  1. Part 1 of 7 · October 19, 2026

    Building Agent Loop: A Local-LLM Agent Runtime in Python

  2. Part 2 of 7 · October 23, 2026

    Why My AI Agent Stopped After 4–5 Iterations

  3. Part 3 of 7 · November 20, 2026

    A Successful Search That Never Searched

  4. Part 4 of 7 · November 27, 2026

    The Crash Window

  5. Part 5 of 7 · December 4, 2026

    When More Evidence Makes an AI Agent Less Reliable

  6. Part 6 of 7 · December 11, 2026

    Evidence-to-Claim Verification

  7. Part 7 of 7 · December 18, 2026

    Testing AI Agents With Independent Agents