Running a one-person company with AI agents: every guide
Seven guides, no signup wall. Each works on its own; together they are the whole rulebook.
- How to Run a One-Person Company with AI Agents (File-Based Setup)Run several AI agent sessions like a small team: one folder as memory, a brief per role, three-line status notes and approval gates. Full setup in 6 steps.
- CLAUDE.md and AGENTS.md Template for a Multi-Agent WorkspaceA CLAUDE.md + AGENTS.md template for running several AI coding agents in one folder: what to put in each, a copyable skeleton, and five mistakes to avoid.
- Human Approval Gates for AI Agents: A 7-Item List for Solo BusinessesWhich actions should an AI agent never take without you? A 7-item approval list for solo businesses, the request format, and keys kept out of files.
- Why Your AI Agent Says "Done" and Nothing ChangedAn agent that reports success without a file has no definition of done. The three-part definition and the status format that enforces it.
- Running Multiple Claude Code Sessions Without ChaosFive sessions, one folder: how to split roles, stop file collisions, keep context small and hand work between sessions without re-explaining.
- Why Agents Ignore Your CLAUDE.md (and Five Fixes)Agents ignore context files for boring reasons: rules buried in prose, contradictions, stale paths, no trigger, no consequence. Five fixes.
- Is a $9.99 AI Company Pack Worth It? What You Get vs FreeAn honest comparison: what the free templates, the 8-file starter and the $9.99 pack each give you, and the one case where you should not buy.
- How to Write a CLAUDE.md (Template and Five Worked Examples)A copyable CLAUDE.md skeleton, five real examples from different kinds of work, and the three lines that make agents stop guessing. Under 700 words.
- AGENTS.md vs CLAUDE.md — Which File Does Your Tool Read?Claude Code reads CLAUDE.md, Codex and others read AGENTS.md. What each is for, how to keep one source of truth, and the two-line setup that avoids drift.
- Three Numbers That Tell You If Your AI Agents Are WorkingStop judging agents by vibes. Three numbers per week — tasks shipped, rework rate, tokens per task — and the one-line sheet that records them.