product-analysis_skill

This skill conducts multi-path parallel product analysis using Claude Code and Codex to generate actionable optimization plans and competitive benchmarks.

Python

609

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill daymade/claude-code-skills --skill product-analysis

SKILL.md9.2 KB

Overview

This skill runs a multi-path parallel product analysis using Claude Code agent teams and optional Codex CLI workers to produce an evidence-backed optimization plan. It spawns parallel explorers across UX, API, architecture, and docs, then synthesizes cross-model findings into prioritized P0/P1/P2 actions. The result is a quantified, actionable audit tailored to the detected project type.

How this skill works

The skill auto-detects available tools (Codex CLI) and the project type, then launches multiple background agents in parallel: several Claude Code Explore agents for different dimensions and, if present, 2–3 Codex CLI invocations for alternative model perspectives. After collection, it cross-validates agent outputs, extracts hard metrics, and generates a structured report with prioritized remediation items and competitive benchmarking when requested.

When to use it

Pre-release product audit to validate readiness and catch usability blockers
Routine self-review for UX, API, architecture, or documentation health
Competitive benchmarking or 'compare' mode to invoke competitor analysis
UX audit and information architecture audits for navigation and onboarding
Fast cross-model validation when you want multiple AI perspectives

Best practices

Run in the repository root so Codex CLI (if present) has full filesystem context
Use 'full' scope for an end-to-end audit; pick 'ux', 'api', or 'arch' to focus
Do not interrupt background agents; allow full synthesis for cross-validation
Treat Codex-only findings as leads to manually verify when disagreement appears
Include a competitor list when using compare mode to get actionable benchmarking

Example use cases

A product team wants a launch checklist highlighting P0 usability defects and API breakages
Engineering lead requests module coupling and dead-code identification before refactor planning
Design and PM run an onboarding audit to reduce steps-to-first-value and quantify interactive density
Security or backend owners request a list of unused or inconsistently handled API endpoints
Benchmark product features and UX against two competitors using the compare workflow

FAQ

The skill proceeds silently with Claude Code agents only; it does not prompt you to install anything.

How are conflicting findings resolved across agents?

The synthesis flags agreement as high-confidence and marks disagreements for targeted re-inspection; Codex-only items are highlighted for manual validation.

Can I limit the analysis to only UX or only API?

Yes. Pass the scope argument (ux, api, arch, full, or compare) to tailor which agents run and which dimensions are analyzed.