One route, one specialist workflow
The router avoids blending several product methods into one vague answer. It selects one skill and then hands control to that skill’s own workflow.
Product Skills is a router for 12 bundled product-team skills. It sends a request to one specialist workflow for prioritization, strategy, UX research, design systems, competitive analysis, product analytics, experiments, discovery, roadmaps, spec-to-repo work, landing pages or SaaS scaffolding. Four related plugins are packaged separately. It is a product-management toolkit, not an ecommerce-specific skill.
Use this router when a product request spans several possible disciplines and the correct workflow is not obvious. It reads the request signals, picks exactly one specialist skill, and asks one clarifying question when two or more routes match. The router itself has no scripts. The selected skill provides its own instructions, scripts, references, and assets.
The router avoids blending several product methods into one vague answer. It selects one skill and then hands control to that skill’s own workflow.
The bundle covers RICE prioritization, OKRs, UX research, design systems, competitive analysis, product analytics, experiments, discovery, roadmaps, spec-to-repo work, landing pages, and SaaS scaffolding.
An ecommerce team can use the workflows for checkout experiments, retention analysis, catalog-tool requirements or roadmap communication. Store data, platform constraints and operating policy still have to be supplied by the team.
The router points to sibling skills. Installing only product-skills/SKILL.md leaves it with nowhere to route the request.
Some selected skills can run scripts or work with repository files. Check the selected skill and keep customer, payment and production credentials out of prompts.
The repository documents this command for installing its skill library into Codex.
npx agent-skills-cli add alirezarezvani/claude-skills --agent codexThe complete source is shown below. Copy it from the top right to use it.
Match the product request to exactly one of these 12 bundled skills: product-manager-toolkit, product-strategist, ux-researcher-designer, ui-design-system, competitive-teardown, product-analytics, experiment-designer, product-discovery, roadmap-communicator, spec-to-repo, landing-page-generator, or saas-scaffolder. Load that skill's SKILL.md and follow its scripts, references, and assets. If two or more routes match, ask one clarifying question before loading anything. If no route matches, say so instead of improvising. This router ships no tools of its own.Starter prompts for the main use cases—copy and use them directly.
Sends a feature-ranking request to the product-manager-toolkit workflow rather than answering with an unstructured opinion. Use when the team has candidate features and enough evidence to compare reach, impact, confidence and effort.
Use Product Skills to route this request: prioritize [FEATURES] for [PRODUCT]. Evidence available: [EVIDENCE]. Business constraint: [CONSTRAINT]. Do not score missing inputs as zero; ask for the single input that would most change the result.[FEATURES][PRODUCT][EVIDENCE][CONSTRAINT]Selects experiment-designer when the team needs a measurable test for a storefront or retention change. Use for checkout, search, recommendation, onboarding or retention hypotheses with a measurable outcome.
Use Product Skills to route an experiment for [CHANGE] on [SURFACE]. Primary outcome: [OUTCOME]. Current baseline: [BASELINE]. Traffic or sample available: [SAMPLE]. Guardrails: [GUARDRAILS].[CHANGE][SURFACE][OUTCOME][BASELINE][SAMPLE][GUARDRAILS]Selects product-analytics when the question requires cohort, funnel or retention evidence. Use when a team needs to diagnose where adoption, checkout or repeat purchase drops.
Use Product Skills to route this analytics question: [QUESTION]. Product event definitions: [EVENTS]. Cohort or segment: [COHORT]. Date range: [DATES]. State which missing field blocks a defensible answer.[QUESTION][EVENTS][COHORT][DATES]Selecting the right product method does not make weak store data, invented customer research or unsupported revenue assumptions valid.
A generated roadmap, experiment or repository scaffold still needs approval from the people responsible for customer impact, data and production systems.
HealthTech CTO and open-source maintainer focused on applied AI, agentic coding, and practical skills for product, research, growth, and operations teams.
Content checked:
No reviews yet.