Bryce Watson
work with me / blog / projects | art / reading / about
Back to blog
Tag · 4 posts

agents

Multi-agent systems: subagent patterns, evaluator agents, harness pruning, and skill architecture.

  • Adversarial Subagent Teams: A Retrospective
    71 invocations of the adversarial-subagent-team pattern in Claude Code and Cowork, verified against the cited PRs, commits, and arXiv references. Across six headline claims: three confirmed clean, one confirmed with caveats, one rejected on falsification, one not supported by the data.
    May 22, 2026 ·
    claude-codecowork
  • Two Patterns That Changed How I Think About Multi-Agent Systems
    Anthropic's latest harness research reveals two underappreciated patterns: why agents can't evaluate their own work (and what to do about it), and why you should be actively removing complexity from your agent harnesses as models improve.
    Mar 29, 2026 ·
    claude-code
  • How to Build Multi-Agent Systems with Claude Code
    Subagents, skills, CLAUDE.md orchestration, headless execution, and state management patterns for Claude Code multi-agent systems. 50 claims verified against official Anthropic docs.
    Mar 16, 2026 ·
    claude-code
  • Case Study: Building a Self-Improving AI Agent That Runs an Etsy Shop
    How I built ShopForge, a production AI agent with three-tier memory, Bayesian confidence scoring, and skill evolution, and what 10 days of real Etsy operations taught me about self-improving systems.
    Mar 08, 2026 ·
    case-study
© 2024–2026 | Bryce Watson