I built an autonomous personal agent — here's the guardrail playbook and the failure that got me
RORDENS · reddit · 2026-10-11
A developer shares hard-won governance lessons from building an autonomous personal agent that connects to email, Calendar and the web, edits Office files, and reports via WhatsApp — with every irreversible action (sending, paying, booking) gated by human approval.
His guardrail checklist, in order of "wish I'd done it earlier":
- Draft, don't send: the agent drafts emails; one unauthorized action resets trust to zero
- Fail-closed: license checked every 15 minutes; unreachable for 72 hours means it stops acting — an agent acting unsupervised is the real risk
- Can't edit its own rules: org-set core rules, since that's the first target of prompt injection
- Hard exclusions: never touches .env files, .ssh folders, password stores
- Audit without liability: logs filter secrets, conversations aren't logged
- WhatsApp as security problem: unknown numbers get nothing but a one-time pairing code; license suspension cuts a number off within a minute
The failure he feared — the agent doing something dumb — never hit. Instead, a stale credential file quietly overrode the long-lived token and the agent went silent for six hours, while the alert meant to catch it was swallowed by its own rate cap. Fixes: pin auth to one token, remove stale files on a schedule, health-check credentials from outside the agent process. Lessons: fail-closed isn't enough, a silent agent must be loud, and governance ended up being the product — users don't ask "how smart is it?", they ask "what happens if it does something dumb?"
More from coding & agent
- Claude Code Pushes Urgent Security Update Across All Installations — umeshonai · 2026-10-23
- Claude cut a 2.6-hour video export to 20 minutes — but only after being told to rebuild the server — byhewar · 2026-10-23
- 3 flight-booking rules my AI agent checks that I'd never do by hand — Efistoffeles · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23