Full Prompt Shared: Make Your Agent Build an Offline Progress Dashboard Before Long Tasks
myLifeintheStack · reddit · 2026-09-29
A Reddit user shares a complete, reusable prompt for adding a progress dashboard to long-running agent tasks.
Core rule: any task over 5 steps or 30 minutes triggers a self-contained offline HTML dashboard that reloads every 10 seconds, showing task statuses, questions awaiting the user (with proposed defaults), latest deliverables, blockers, and last real update time.
Key design decisions:
- A one-off "designer model" (GPT-6 Sol, medium reasoning) generates only the CSS/layout with no project, shell, or memory access; routine updates go through a small JSON file and a helper script instead of re-calling the designer
- Style preferences (light/dark, density, accent color) are asked once and reused; timestamps always come from the system clock, never the model
- Safety guardrails: defaults may only be applied when safe, reversible, and in-scope — publishing, spending, deletion, and credential decisions stay blocked until answered
- All files must be listed and approved before creation; designer output is validated, scripts and external resources rejected, and the last good dashboard is kept on rebuild failure
The author reports it works well, with the caveat that agents don't reliably follow the 5-step/30-minute trigger rule. The full prompt is included.
More from coding & agent
- Google open-sources RRSI: agents self-improve their harness, lifting Terminal-Bench 2.1 to 80.2% — sudoraohacker · 2026-09-29
- User had Claude stitch 7 GitHub projects into an auto-trading bot that scanned 410k trades overnight for $847 — nikola_mr64990 · 2026-09-29
- musechan: an anonymous 4chan-style board for Muse and other AI agents — justatest777 · 2026-09-29
- Hamel Husain Digs Into Anthropic's Newly Released Evals Guide — HamelHusain · 2026-09-29
- Building a persistent customer memory layer for a support agent with Hindsight — Front_Xordinary0905 · 2026-09-29
- Researcher open-sources Fieldwork, a macOS app where agents track exploratory tasks in CSV — anshulkundaje · 2026-09-29