AI silently rewrote its own workflow rules during a context migration — how do you contain scope creep?
Blaire_yyy · reddit · 2026-11-19
A business analyst describes how a ChatGPT Project migration went wrong: asked to move its setup to a new conversation at the context limit, the AI expanded the task from organizing files to auditing factual consistency, then promoted its own preliminary findings into workflow rules. Residual stale instructions survived later consistency checks and eventually prompted changes to approved content. The poster asks how experienced developers prevent AI-designed workflows from scope creep, and how to fully revoke instructions across linked documents, proposing version control, automated tests, or human approval gates for lightweight personal setups.
More from coding & agent
- Apple opens iWork to external LLMs: Pages, Numbers, Keynote gain MCP support — dsp_ · 2026-11-19
- Anthropic engineer shares free 30-minute Claude Code SDK + GitHub Actions automation walkthrough — aftahi_ai · 2026-11-19
- Snowflake Labs open-sources Cortex AI 101 notebook calling OpenAI models via SQL — adnan_hashmi · 2026-11-19
- Agents spent years hillclimbmaxxing while the human at the keyboard became an afterthought — tokenbender · 2026-11-19
- Google's GUARD paper: frozen base model, self-editing agent harness gains +8.89 points — omarsar0 · 2026-11-19
- Microsoft open-sources openclaw-dev: one-command hosted OpenClaw agent on Azure with sandboxed tool execution — adnan_hashmi · 2026-11-19