Claude Opus 5 prompt tips say old harness habits now waste tokens
tengyanAI · x · 2026-07-25
A user shares prompting notes for Claude Opus 5 after testing it in agent-style workflows.
Their main takeaways: avoid redundant instructions like asking the model to verify or double-check, because it may already do that internally; don’t spawn subagents just to review its own work; and if you want concise output, say so explicitly. The post also suggests giving the model the full task up front and letting it run, since effort controls reasoning rather than verbosity.
More from coding & agent
- img2threejs is being used to give AI coding agents procedural 3D generation — anselm · 2026-07-25
- Codex and Three.js drive an interactive procedural house demo — anselm · 2026-07-25
- T3 Code ships 76 PRs of updates, including Opus 5 support and auto approvals — TAbrodi · 2026-07-25
- A coding agent runs two commands, then saves a fact to memory — sull · 2026-07-25
- A coding-agent dashboard tracks 1.9B tokens, $2,595.89 in spend, and 417 chats — zeeg · 2026-07-25
- InstructSAM gets a C++/GGUF port for local multimodal segmentation with Qwen3-VL-2B — remyxai · 2026-07-25