anupsahoo

Revenueos — AI skill for Claude Code

AI community

RevenueOS — an event-sourced Sales→PreSales handoff: an AI agent drafts an editable POC plan from a reusable template library, a person decides, and the seam watches its own SLA.

How to install Revenueos

This entry records only its repository, not the path inside it, so there is no exact command to give. Open anupsahoo/revenueos and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Revenueos does

RevenueOS — an event-sourced Sales→PreSales handoff: an AI agent drafts an editable POC plan from a reusable template library, a person decides, and the seam watches its own SLA. Next.js + TypeScript + Claude.

Alternatives in AI

LLM For Zotero — An open-sourced research agent system deeply rooted in your Zotero library 2.8k ★
UX Lint — Deterministic regex-based linter for AI fingerprints 66 ★
Next Chat Skills — AI assistant application based on Next.js, supporting Dialog Management and external Skill (script/tool/rule) 57 ★

README

RevenueOS — Brief→POC-Plan Loop

An operating system for a revenue handoff. When Sales wins a deal, the Sales→PreSales handoff (the "seam") is cleared by an AI agent, a person decides, the system learns from that decision, and the seam watches its own SLA health.

The loop

Won opportunity  ─▶  OS event (brief)  ─▶  Agent retrieves templates + drafts
   (CRM-shaped)                              POC plan & handoff skeleton
                                                     │
                                                     ▼
   Seam health  ◀──  OS events  ◀──  Solution Architect accepts / edits / rejects
   (age vs SLA,                             (decision is itself an event;
    green/amber/red,                         reuse rate + template ranking learn)
    trigger on breach)

The problem it solves

A B2B platform sells to banks and insurers. Deals stall at one seam: when a brief reaches PreSales, a solutions engineer must pick it up and start the POC plan, but prior solution templates aren't retrievable at that moment — so every POC is built from scratch and briefs sit idle against their SLA.

This loop makes those templates **retrievable, explainable, and reusable** at the instant a brief arrives, and measures reuse going up.

Principles

Every handoff is an event — timestamped and owned; the event log is the source of truth.
Health is computed, not reported — SLA state is derived from events, never typed in.
Agents do the volume work; people decide — the agent drafts and explains; a person commits.

Stack

Next.js (App Router) + TypeScript, deployed on Vercel — one repo, one control surface, API routes.
Anthropic Claude agent (tool-calling) for retrieval-explanation + drafting.
Append-only event store behind one interface — in-memory + seed for demo/local (no accounts to run), swappable to Neon / Vercel Postgres for persistence.
Explainable retrieval over a seeded library that **lea