innovation-research

name	innovation-research
description	Run a structured innovation research cycle — discover gaps, propose novel capabilities, prototype and validate breakthrough ideas for structural engineering.
argument-hint	Domain: 'sustainability' | 'generative-design' | 'ml-estimator' | 'failure-analysis' | 'knowledge-graph' | 'cross-code' | 'digital-twin' | 'construction-intel'

Skill: Innovation Research

Run a structured innovation research cycle to discover, validate, and prototype new capabilities for the structural engineering library.

When to Use

• Starting a research session on a new innovation domain
• Evaluating feasibility of a proposed innovation
• Prototyping a new capability
• Reviewing the innovation backlog
• Surveying what competitors and open-source tools offer

Pre-Flight Checks

# 1. Check existing research docs
ls docs/research/

# 2. Check library coverage gaps
.venv/bin/python scripts/parity_dashboard.py

# 3. Check existing insights/AI features
ls Python/structural_lib/insights/
grep "^def " Python/structural_lib/insights/*.py 2>/dev/null

# 4. Check current API surface
grep "^def " Python/structural_lib/services/api.py | head -20

# 5. Check innovation backlog in TASKS
grep -i "innovation\|research\|novel\|prototype" docs/TASKS.md

Step-by-Step Cycle

Step 1: Scan — Survey the Landscape

Review the Innovation Domains table in .github/agents/innovator.agent.md. Check what exists in our codebase and what competitors offer.

# What do we already have?
ls Python/structural_lib/codes/is456/    # Implemented IS 456 modules
ls Python/structural_lib/insights/       # Existing smart features
ls Python/structural_lib/services/       # Service layer

# Use web search for competitor analysis (innovator has web access)

Step 2: Identify — Find the Gap

Pick the highest-impact innovation domain that's still in "🆕 Research" status. Consider:

• Engineering value (does this solve a real problem?)
• Uniqueness (does any other tool do this well?)
• Feasibility (can we build it with our stack?)

Step 3: Research — Deep Dive

Use web search to find:

• Academic papers on the topic
• Open-source implementations (OpenSees, SAP2000 API, etc.)
• Industry standards and databases
• Benchmark data for validation

Document findings as you go — research without docs is lost.

Step 4: Propose — Write Innovation Document

Create docs/research/innovation-<domain>.md using the Innovation Proposal Template from innovator.agent.md.

Required sections:

• Problem Statement
• Current State (what exists)
• Proposed Approach
• Data Requirements
• Implementation Sketch
• Validation Plan
• Impact Assessment

Score feasibility using the 7-criterion framework (must be ≥40/70 to proceed):

Criterion	Score (1-10)
Engineering Value	?
Uniqueness	?
Feasibility	?
Data Availability	?
User Demand	?
Code Integration	?
Safety	?
Total	?/70

Step 5: Prototype — Build Proof of Concept

Create prototype in Python/structural_lib/research/ (create directory if needed). File naming: research_<domain>.py

Delegate implementation to specialist agents:

• @structural-math — math prototypes, core types
• @api-developer — API endpoints
• @frontend — UI/visualizations
• @tester — validation tests, benchmarks

Step 6: Validate — Engineering Validation

Hand off to:

• @structural-engineer — correctness and IS 456 compliance check
• @library-expert — professional standards review
• @security — safety and security implications

Run benchmarks against known results where applicable.

Output

A complete innovation cycle produces:

• Innovation proposal in docs/research/ (with feasibility score)
• Prototype code in Python/structural_lib/research/ (if feasible)
• Validation results documented
• Delegation plan for production implementation
• TASKS.md updated with any new innovation tasks
• next-session-brief.md updated with research continuity

Example Usage

# Research sustainability scoring domain
1. Check existing: ls Python/structural_lib/insights/
2. Web search: ICE database, embodied carbon calculation methods
3. Write proposal: docs/research/innovation-sustainability-scoring.md
4. Score feasibility: Engineering Value=9, Uniqueness=8, ...
5. Prototype: Python/structural_lib/research/research_sustainability.py
6. Validate: delegate to @structural-engineer

Related Skills

• /api-discovery — check existing API before proposing new features
• /is456-verification — validate IS 456 compliance of prototypes
• /new-structural-element — if innovation requires a new element type
• /architecture-check — verify prototype fits 4-layer architecture