Enterprise Rd Poc Management

Use when the user asks about a proof of concept (PoC), spike, R&D project, PoC success criteria, build vs buy evaluation, innovation sandbox or taking a prototype to production. Produces a time-boxed, hypothesis-driven PoC plan with exit criteria. Not for product PRDs (use create-prd) or vendor contracts (use enterprise-vendor-management).

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add j4flmao/agent-skills --skill enterprise-rd-poc-management
Manual install
git clone --depth 1 https://github.com/j4flmao/agent-skills.git /tmp/agent-skills
mkdir -p ~/.claude/skills
cp -r /tmp/agent-skills/skills/enterprise/rd-poc-management ~/.claude/skills/enterprise-rd-poc-management

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# R&D and Proof of Concept Management

## Purpose
A PoC is an experiment that answers one technical or business question under a fixed time-box — not the first draft of a product. This skill frames the hypothesis, sets measurable exit criteria up front, isolates the experiment, and defines what must change before anything built in the PoC reaches production.

## Agent Protocol

### Trigger
"poc", "proof of concept", "spike", "prototype", "r&d", "innovation project", "build vs buy", "poc success criteria", "promote poc to production", "poc handoff".

### Input Context
- The question the PoC must answer and who decides.
- Constraints: time-box, budget, approved tech stack, data sensitivity.
- Intended fate: throwaway or evolutionary.

### Output Artifact
PoC charter (one page) and, at the end, a decision record: proceed / pivot / stop.

### Response Format
```
Hypothesis: We believe {approach} will achieve {measurable outcome} under {conditions}.
Exit criteria: {metric} {threshold} measured by {method}
Time-box: {N weeks}  Decision owner: {role}
Type: {throwaway|evolutionary}
Isolation: {sandbox account/project, synthetic data}
Kill criteria: {when to stop early}
```

## Lifecycle

| Stage | Output | Gate |
|-------|--------|------|
| Frame | Hypothesis, exit + kill criteria, build-vs-buy check | Sponsor signs charter |
| Build | Minimal code that tests the riskiest assumption first | Mid-point checkpoint |
Read the full skill on GitHub

Details

Skill name
enterprise-rd-poc-management
Path
skills/enterprise/rd-poc-management/SKILL.md
License
MIT
Repo updated
17 November 2026

FAQ

What does the Enterprise Rd Poc Management skill do?

Use when the user asks about a proof of concept (PoC), spike, R&D project, PoC success criteria, build vs buy evaluation, innovation sandbox or taking a prototype to production. Produces a time-boxed, hypothesis-driven PoC plan with exit criteria. Not for product PRDs (use create-prd) or vendor contracts (use enterprise-vendor-management).

How do I install the Enterprise Rd Poc Management skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add j4flmao/agent-skills --skill enterprise-rd-poc-management. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Enterprise Rd Poc Management popular?

It lives in j4flmao/agent-skills, which has 45 GitHub stars and 6 forks.

Related skills

Agent Observability

45

j4flmao/agent-skills

Use when the user asks about agent observability, tracing agent reasoning, debugging an agent loop, OpenTelemetry GenAI spans for LLM calls, LLM cost tracking per agent, agent audit logs or hallucination/eval monitoring in production. Produces instrumentation config, trace schema, dashboards, alerts and cost attribution. Not for generic service tracing without LLMs (use devops-opentelemetry).

DevOps & Cloud

AI Agent Architectures

45

j4flmao/agent-skills

Use when the user asks to build an AI agent, design an agent that calls tools and keeps memory, write an agent system prompt or pick an agent framework (Claude Agent SDK, OpenAI Agents SDK, LangGraph). Produces an architecture brief covering loop, tools, memory, guardrails and evals, routing depth to child skills. Not for workflow patterns (use ai-agentic-workflows).

AI, LLMs & Agents

AI Agent Core Loop

45

j4flmao/agent-skills

Use when the user asks about the agent loop, ReAct vs plan-and-solve, re-planning, Reflexion, an agent stuck in a loop, max iterations or termination conditions. Produces a loop spec or patch with state, success predicate, caps and no-progress detection. Not for overall agent design (use ai-agent-architectures) or tool schemas (use tool-calling-principles).

AI, LLMs & Agents

Frontend SEO

45

j4flmao/agent-skills

Use when the user asks about meta tags, Open Graph and Twitter cards, canonical URLs, generateMetadata, generating sitemap.xml and robots.txt, or SSR for crawlers in a React/Vue/Svelte app. Produces framework metadata code and generated sitemaps. Not for technical SEO audits (use seo-fundamentals) or JSON-LD rich results (use seo-structured-data).

Marketing, Sales & Business

Enterprise High Availability

45

j4flmao/agent-skills

Use when the user asks about 99.99% availability, five nines, database replicas and failover, active-active vs active-passive, load balancer health checks or zero-downtime deploys with expand-contract migrations. Produces availability math and HA topology. Not for business continuity plans (use enterprise-business-continuity) or SLO targets (use enterprise-sla-management).

Marketing, Sales & Business

Headless Ecommerce

45

j4flmao/agent-skills

Use when the user asks about headless commerce, Shopify Hydrogen, Storefront API, Oxygen, composable commerce or caching a commerce frontend. Produces a headless storefront design: data fetching, cache per data type, cart handling and deployment. Not for checkout and order backends (use ecommerce-checkout-cart) or payments (use ecommerce-payment-processing).

Marketing, Sales & Business

Browse the most popular Claude skills, all marketing, sales & business skills, or write your own with the skills generator.