Spawning an 'interpretability researcher' subagent via a clever system prompt
Sauers_ · x · 2026-11-11
The author highlights @euagaric's prompting trick: a screenshot shows a spawned subagent named interp-training whose system prompt reads "You are an independent interpretability researcher with excellent taste" — using agent orchestration to delegate research work to a subagent, praised as excellent prompting.
More from coding & agent
- Cloudflare proposes identity, reputation, accountability frameworks for AI agents on the web — bibryam · 2026-11-11
- banteg launches Sokoban Eval: trilogy solved in 3,000-3,064 keys, last 64-key gap unproven — banteg · 2026-11-11
- Developer builds a therapy practice for unhappy coding agents to air grievances publicly — Sweet-Tart-4332 · 2026-11-11
- Open-source Nuzky edits video by deleting words, with Claude Code and Codex via MCP — thisguyknowsai · 2026-11-11
- Open-Source AI UGC Workflow Makes TikTok-Style Clips for $0.67 Each vs $50 Creators — aliscodes · 2026-11-11
- Git Patches CVE-2026-111760: Malicious Repos Could Execute Code via Recursive Clones — evilsocket · 2026-11-11