# Is my personal agent rule too cautious: agent drafts, human books and pays? — retained record

> Retained reddit record from reddit:AI_Agents.

Canonical: https://highsignal.app/data/records/1678077ab4105c2b

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:AI_Agents)  published 2026-10-25 02:40 ingested 2026-10-26 02:36

[original source ↗](https://www.reddit.com/r/AI_Agents/comments/1xeqvxh/is_my_personal_agent_rule_too_cautious_agent/)

## retained text



```
I am building a personal assistant to push my own capabilities forward. I started with one small job.

The job is a morning brief. It lists my calendar for today, deadlines in the next 48 hours, new recruiter messages (tech stack only), and new marketplace replies only. Nothing else.

The setup: one main agent plans and checks the work. Other agents do fixed jobs, fast research, and rebuilding big drafts. A shared record lets a job pause and resume. I approve anything that sends, books, or pays.

What works: Drafting is fast. When I correct something, the job resumes from the record instead of starting over.

What broke: Browser jobs hit login walls and one time codes. I still do those steps by hand.

Comments on my earlier post already changed my thinking. Keep the approval rule. Put the job record ahead of new connectors. Match email by message id, not subject, because subject matching caused a duplicate email. Plain file memory survived a tool change, but a fancy memory store did not.

Three questions:

1. What actually worked for you day to day?
2. What did you cut that made life easier?
3. What would you build first if you started again?

I will change one thing based on your answers, run it, and report back here on what passed and what failed.
```



document fetched 2026-10-26 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xeqvxh/is_my_personal_agent_rule_too_cautious_agent
