Building a personal chief-of-staff agent: should humans approve every send, booking and payment?
Affectionate-Sun265 · reddit · 2026-10-25
A developer is publicly building a personal chief-of-staff agent and asking the community to validate one core design rule: the agent (Muse orchestrating, Zapier for fixed jobs, Gemini for research, Claude for major drafts, state in Sheets/Drive) researches, drafts and monitors, but a human approves anything that sends, books or pays.
He worries the rule may be too cautious to be useful and asks three questions he'll act on: what first workflow passed twice with no duplicate action; where others draw the line on send/book/pay and what broke when they crossed it; and what memory setup survived a framework change. He commits to shipping one change and posting what passed or failed.
More from coding & agent
- Slop Eisley: a Star Wars fan game fully coded by Claude Opus and Codex opens beta with public invite codes — jasonkneen · 2026-10-25
- Weekend opsec: hand Claude a checklist to audit your home network and devices — zedlander · 2026-10-25
- Cornell paper: 64.7% of AI agent rollouts melt down on errors, no adversary needed — 2C_ornot2C · 2026-10-25
- Blind test of 5 local AI agents on one Qwen model: scores range from 3/5 to 5/5 — GapNew4766 · 2026-10-25
- WildAWM synthesizes open-source computer-use agent worlds; Qwen3.6 gains +10.6 across 8 benchmarks — HuaxiuYaoML · 2026-10-25
- Meme: Claude as Sisyphus endlessly pushing the 'Some Rust Port' boulder — burny_tech · 2026-10-25