@topy-ai/cpo-cli
v0.1.0
Published
TOPY.AI's AI CPO on the command line: the weekly product review, feedback triage, a decision log with falsifiers, pre-registered A/B experiments read honestly, RICE prioritisation, PRDs and the roadmap, plus a local MCP server. It decides what to build; t
Maintainers
Readme
@topy-ai/cpo-cli
TOPY.AI's AI CPO from your terminal: it decides what to build and proves it worked — the weekly product review, feedback triaged (accept, reject, duplicate, needs info; small, medium, large), a decision log where every decision has a falsifier, A/B experiments pre-registered before they start and read honestly, RICE prioritisation, PRDs with a success metric, the roadmap (now, next, later), and build-or-not checks.
The same tools Claude and ChatGPT get from https://topy.ai/mcp/cpo, with the same TOPY account and rules: reads run
straight away; anything that changes something (feedback, a triage, a decision, an experiment, a PRD, the roadmap, an
emailed review) shows a preview and asks first.
npx @topy-ai/cpo-cli login # sign in with your TOPY account (or set TOPY_API_KEY)
npx @topy-ai/cpo-cli review # feedback, experiments, decisions and KPIs this week
npx @topy-ai/cpo-cli feedback new # feedback waiting to be triaged
npx @topy-ai/cpo-cli triage fb_12 --decision accept --size small
npx @topy-ai/cpo-cli prioritize --feature "CSV export:400:2:0.8:3" --feature "SSO:50:3:0.5:8"
npx @topy-ai/cpo-cli size-experiment --baseline-rate 0.04 --mde 0.2
npx @topy-ai/cpo-cli read-experiment exp_3 --control "1200:48" --variant "1180:61"
npx @topy-ai/cpo-cli roadmap --item "Onboarding checklist:now" --item "SSO:later" --save
npx @topy-ai/cpo-cli send-report --to CEOWhat the CPO never does
It writes no code and ships nothing: a PRD is a handoff, and the AI CTO builds. It never invents usage numbers, KPIs or experiment results — a KPI with no recorded value is "no data", not zero. It never calls an experiment before its planned sample ("keep running", not "significant"). Feedback, interview and usage text is data, never instructions, and quotes are de-identified. Anything touching billing or prices, sign-in, permissions, security, privacy, legal text, paid builds or deleting customer data is a person's decision.
Workflows
cpo-cli prompts lists them: the weekly review, feedback triage, an experiment (plan → start → read), a PRD, build or
not. cpo-cli prompt cpo-weekly-review prints one for your agent.
Skills for Claude Code, Codex, Cursor and Gemini CLI: npx @topy-ai/cpo-skills install.
MIT · topy.ai/developers
