Employee Hub — AI employees for any business
Five new AI employees every morning at 06:00 IST. One folder per employee, one skill file that carries the craft, and a written standard for how every employee behaves when it cannot do the job.
What it is
A person you hire brings four things: experience, tool access, a brief about your business, and a schedule. Employee Hub treats an AI employee as those same four parts — the skill file carries the experience, the connectors carry the tools, my context files carry the brief, and a cron run or a webhook carries the schedule.
The part that matters most
Most skill collections stop at the skill. This one adds a standard every employee follows, and it has four obligations: check the connectors before answering, so no employee produces a report it had no way to compute; keep notes, so it improves instead of repeating itself; track itself on the calendar, with a weekly review and a monthly scorecard; and never fake the part it cannot do — deliver what it can, name the gap, offer a manual fallback.
That last obligation is the whole reason the project exists. An AI employee that invents a market price is worse than no employee at all.
What is actually inside
- 716 files: 687 markdown documents, 57 of them
SKILL.mdfiles, and 21 Python scripts that do the deterministic work so the model does not have to reason about it. - A plugin marketplace manifest, a changelog, a contributing guide and an MIT licence, because it is meant to be installed and extended, not admired.
- Connector preflight documents and a connector catalogue, so an employee knows what it can reach before it starts.
- An audit document and a daily-drop schedule behind the five-a-morning cadence.
What it taught me
A skill is easy. A standard is hard. The moment you have more than one employee, the behaviour when something is missing matters more than the quality of the output when everything is present — because the missing case is the one nobody checks.