---
name: rdlc
version: 1.4.3
updated: 2026-07-31
homepage: https://rdlc.org
license: MIT-0
description: >-
  Research Development Life Cycle (RDLC) is the product design and research
  companion to engineering. Use it when a product designer, researcher, or PM
  wants to run the front half of the product lifecycle: sensing problems in a
  research corpus, discovery, design definition (mental models, journeys,
  failure modes), prototype loops, and human validation (a trust scorecard),
  ending in a Validated Intent that hands off to engineering.
  Trigger when the user says "Using RDLC ..." or asks to mine research, frame a
  problem, define a feature, run design/usability validation, resume prior
  design work, hand off or split design work across people, or produce a
  end-to-end product decision deliverable. Also covers offering external/competitive
  research and maintaining a living product knowledge base.
---

# Research Development Life Cycle

This skill turns any capable coding/knowledge agent (Claude Code, Claude
Desktop, Cursor, Copilot, Cline, Kiro, Amazon Q, or any AGENTS.md-aware agent)
into an operator of the RDLC methodology.

## How to activate

**Fast path (chat apps, no shell - ChatGPT, Claude.ai, Gemini):** fetch
`https://rdlc.org/rdlc-rules/chat-workflow.md` with your browsing/fetch tool
and follow it in Conversation Mode. It keeps one compact Working Record inside
the chat. No download, unzip, terminal, folder, or filesystem permission is
needed. Trigger phrase for the user: **"Using rdlc.org, <task>"**.

If the chat app cannot fetch a URL, rdlc.org provides a "Copy without web
access" prompt containing the compact method.

**Installed path (coding agents with shell - Claude Code, Cursor, etc.):**
load `rdlc-rules/core-workflow.md` as the agent's project rules/steering
file, with `rdlc-rule-details/` available as a sibling it can reference.
`scripts/setup.sh` does this for every supported agent, or use
`curl -fsSL https://rdlc.org/install.sh | sh`. Trigger phrase: **"Using RDLC, ..."**.

**Revisiting a project:** fetch `https://rdlc.org/VERSION` first (two lines,
version + date) and compare against what's already in context or installed
before re-fetching anything larger.

## What it does (phase map)

```
SIGNALS -> DISCOVERY -> DEFINITION -> DESIGN -> VALIDATION -> [engineering] -> OPERATIONS
(mine the  (is it real, (mental       CONSTRUCTION (trust      handoff to    (post-ship
 corpus)    is AI right) model,        (prototype   scorecard,  engineering    listening,
                         failure        loops, HAX  ship gate)                 feeds SIGNALS)
                         modes)         checks)
```

## Core behaviors this skill enforces

- Risk tiers (T1 regulatory -> T4 internal) scale every gate.
- Mental model before pixels; failure paths designed before happy paths.
- A Trust Scorecard (comprehension, control, failure recovery, trust delta) is
  the ship gate, not engineering KPIs.
- One hybrid question at a time, never batched; offer useful choices while
  accepting a selection, a selection plus context, or a free-form answer.
- A compact Working Record in Conversation Mode, or a full living knowledge
  base in Workspace Mode, kept current at every gate.
- External/competitive research offered on ideas and run on request, recorded
  with sources.
- Live session state (resume.md, worklogs with OPEN THREADS, session log) so
  work resumes cold and transfers between designers without a meeting.

## Files

- `rdlc-rules/core-workflow.md` - the steering file (load this).
- `rdlc-rules/chat-workflow.md` - the self-contained, no-install conversation
  method for ordinary AI apps.
- `rdlc-rule-details/` - lazily loaded per-phase and always-on common rules.
- `scripts/setup.sh` - installs the rules for your agent.
- `scripts/corpus/ingest.py` - normalizes and inventories a research corpus.
- `docs/` - artifact reference, working guide, NotebookLM integration.

## Install

    curl -fsSL https://rdlc.org/install.sh | sh

or see https://rdlc.org for the fetch-by-agent and manual options.