Sakhanda Wire
NVDA $152.56 -2.29% MSFT $505.79 -2.36% GOOGL $279.70 -1.42% META $618.14 -1.32% AMZN $256.03 -0.67%
← Back to the news

GPT-6 Astra needs shorter prompts and fewer restrictions

OpenAI recommends simplifying prompts and project instructions when developers move to GPT-6 Astra. Eric Provencher says accumulated skill descriptions, broad reading requirements and rigid approval rules can consume context or make the model stop before the task is complete. He advises reviewing skills, AGENTS.md and task prompts after a model change.

Skills are Markdown-based instructions that help Codex select workflows. Provencher recommends giving them narrow, precise descriptions and linking to supplementary documents only when needed. Too many or overlapping skills can truncate the information used for selection, while detailed step-by-step procedures may unnecessarily constrain a model that handles ambiguity better.

The same approach applies to repository rules. Documents should be required selectively—for example, architecture material for service-boundary work and database documentation for schema changes. Safe local testing can be explicitly permitted to avoid repeated confirmations. Provencher also says teams should define what “done” means, including implementation, testing and error correction, rather than relying on restrictive checkpoints. The guidance builds on OpenAI’s earlier prompting recommendations for GPT-6 Astra.

Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.

Originally published by The Decoder on

Read the original on The Decoder ↗

The reporting and the image belong to The Decoder and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news