om-discover

Guides product discovery interviews and writes an evidence-tagged product-brief.md.

1|Updated Apr 23, 2026
One-click install
npx skills add https://github.com/tkogut/agents-os-core --skill om-discover-tkogut
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: om-discover
Source: https://github.com/tkogut/agents-os-core/tree/main/global_skills/om-discover
Command: npx skills add https://github.com/tkogut/agents-os-core --skill om-discover-tkogut

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Product teams often start building on untested assumptions, invented research, or vague stakeholder requests. This Skill structures a product discovery conversation that separates observations, decisions, and hypotheses, then records the result in a rigorous product brief that downstream skills can trust. ## Core Features & Use Cases - Structured discovery interviews: Runs bounded question rounds in three modes (existing product, client idea, own idea), asking only questions that could change the current decision. - Evidence-disciplined brief: Writes product-brief.md with source-tagged claims ([INTERVIEW], [DATA], [DOCUMENT], [PRODUCT], [BENCHMARK], [SYNTHETIC], [ASSUMPTION]), stable decision/rule/non-goal ids, and a collection plan for missing material. - Independent skeptical review: Passes the draft to a fresh-context subagent that checks source fidelity, problem/solution fit, and readiness before anything is written. - Use Case: A founder with a new SaaS idea runs a discovery session; the Skill interviews them, tags every claim by evidence tier, records confirmed decisions with owners, and produces a brief that honestly states which assumptions remain untested. ## Quick Start Ask the assistant to run a product discovery session for your idea or existing product and produce a product brief.

Frequently Asked Questions about om-discover

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a product discovery session with an AI assistant?▼

Invoke the skill with an optional topic and mode flag, such as --mode existing, client, or own. It reads available research and repository material first, then asks two or three decision-relevant questions per round before drafting the brief.

What is a product brief and what should it contain?▼

A product brief records the problem with evidence, target users, scope, business rules, non-goals, decisions, riskiest assumptions, and open questions. This skill writes it to product-brief.md with stable ids and source tags so other skills can consume it.

How does the skill prevent invented user research or fake data?▼

A hard gate forbids fabricating research, quotes, or measurements. Every claim carries an evidence tag like [INTERVIEW], [DATA], or [ASSUMPTION], and a skeptic review checks each claim against its cited source before the brief is written.

Can I use it for a quick discovery pass instead of a full session?▼

Yes, the --quick flag runs one question round using already available material, with the skeptic checks applied inline. It discloses the lack of independent review and defers details irrelevant to the current decision.

Does the discovery skill work without a configured tracker or pipeline?▼

Yes, configuration is optional. Without it the skill uses .ai/specs as the specs directory and skips read-only tracker operations, while the full discovery conversation and brief writing still work.

When should I not use a synthetic user panel after discovery?▼

A synthetic panel requires a concrete Key flow in the brief and a reason it would inform the current decision. Without a flow the handoff is skipped, and synthetic reactions are always tagged as hypotheses, never as evidence of demand.