Quick Reference Card

Print this page (Ctrl+P → Save as PDF or print double-sided). Optimized for A4 paper and narrow screens.

---

Key Actions at a Glance

1. Before the event → Run the Participation Gate
2. First 10 minutes → Open the Dev Container, sign in only where needed, and verify current agents, models, and MCP servers
3. Each challenge → Check inputs, use the owning main agent, resolve required reviews, approve the handoff, and preserve evidence
4. End of day → Team lead deletes resources and confirms cleanup

---

Hackathon Schedule (1 Day)

Time	Challenge	Duration	Points
09:00-10:00	Intro	60 min	—
10:00-11:00	Challenge 1	60 min	20
11:00-12:00	Challenge 2	60 min	25
12:00-12:45	🍽️ Lunch	45 min	—
12:45-13:30	Challenge 3	45 min	25
13:30-14:15	Challenge 4	45 min	10
14:15-14:30	Checkpoint	15 min	—
14:30-15:00	Challenge 5	30 min	5
15:00-15:15	☕ Break	15 min	—
15:15-15:30	Challenge 6	15 min	5
15:30-15:35	Challenge 7	5 min	5
15:35-15:50	Prep	15 min	—
15:50-16:50	Challenge 8	60 min	10
16:50-17:00	Wrap-up	10 min	—

Total Points: 105 base + 25 bonus

Scoring & Leaderboard

Method	How
Rubric scoring	All scoring is manual. Facilitators review artifacts and verify deployments against the rubric.
Leaderboard	Use your facilitator worksheet or HackerBoard instance if your event package includes one
Tooling	Optional: record rubric-based totals in HackerBoard when that tooling is available (it does not calculate scores automatically)
Rubric	Scoring Rubric (available from facilitator) — single source of truth for points

---

Keyboard Shortcuts

Shortcut	Action
Ctrl+Alt+I	Open Chat view
Ctrl+Shift+I	Switch to Agent mode
Ctrl+I	Inline chat (editor/terminal)
Ctrl+N	New chat session
Ctrl+Alt+.	Model picker
Tab	Accept suggestion
Escape	Dismiss suggestion

---

Custom Agents

Select main agents from the agent dropdown. The human selects and approves each main-agent handoff.

Agent	Purpose	Challenges
01-Orchestrator	Identify the next step and propose a human handoff	All, optional entry point
02-Requirements	Capture requirements and initial SKU evidence	1, 4 revision
03-Architect	Assess architecture and cost feasibility	2, 4 revision
04-Design	Create optional diagrams and ADRs	2, 4
04g-Governance	Discover and reconcile effective Azure Policy constraints	3, 4 revision
05-IaC Planner	Create the implementation plan and contracts	3, 4 revision
06b-Bicep CodeGen	Generate the approved Bicep track	3, 4
06t-Terraform CodeGen	Generate the approved Terraform track	3, 4
07b-Bicep Deploy	Deploy the authorized Bicep scope	3, 4
07t-Terraform Deploy	Deploy the authorized Terraform scope	3, 4
08-As-Built	Generate evidence-based as-built documentation	6
09-Diagnose	Diagnose the selected scope from current evidence	7
10-Challenger	Review requirements, architecture, cost, governance, and plan artifacts	1-4
11-Context Optimizer	Audit Copilot context usage and recommend improvements; reference-only	—

Check the current .github/agents/*.agent.md files for model selections. Do not rely on a copied fixed model list. The Context Optimizer is reference-only and does not replace an owning agent or a human approval gate.

How to use: Ctrl+Alt+I → select the owning agent → provide current artifacts and scope → review output → resolve findings → approve the next handoff.

---

Chat Features

Feature	How to Use
Context	#file, #folder, #symbol, drag & drop
Codebase search	#codebase in prompt
Fetch web page	#fetch url
Workspace	@workspace for workspace questions
Terminal	@terminal for shell help
Slash commands	/fix, /explain, /tests, /doc

---

Essential CLI Commands

Participant deployment is agent-owned. Use only the approved 07b-Bicep Deploy or 07t-Terraform Deploy agent after the team reviews current validation and explicitly authorizes the proposed scope. Do not run Azure deployment commands directly as a participant. The CLI examples below are facilitator diagnostic/evidence examples only.

az account show --query name -o tsv
bicep --version

# Create resource group
az group create -n rg-freshconnect-dev-swc -l swedencentral

# Validate Bicep
bicep build main.bicep
bicep lint main.bicep

# Facilitator diagnostic example: What-If
az deployment group what-if -g rg-freshconnect-dev-swc -f main.bicep

# Facilitator diagnostic example: direct deployment evidence
az deployment group create -g rg-freshconnect-dev-swc -f main.bicep

# Cleanup (END OF DAY — team lead is responsible!)
az group delete -n rg-freshconnect-dev-swc --yes --no-wait
az group delete -n rg-freshconnect-dev-gwc --yes --no-wait  # if secondary region was used
# Verify cleanup:
az group list --query "[?starts_with(name, 'rg-freshconnect')]" -o table

Ask your facilitator to remove governance policies from the team subscription.

---

Bonus Targets (+25 max)

Enhancement	Points	What to prove
Zone Redundancy	+5	Supported SKU with zone redundancy enabled
Private Endpoints	+5	Private connectivity for the relevant services
Multi-Region DR	+10	Real DR design or deployment spanning 2+ regions
Managed Identities	+5	Identity-based access with no hard-coded connection strings

---

Expected Outputs

Challenge	Input evidence	Output evidence	Next action
1	Scenario brief	Requirements, SKU manifest, requirements review findings, approval	C2: Architecture
2	Approved C1 evidence	Architecture, cost evidence, separate reviews, approval, optional design artifacts	C3: Implementation
3	Approved C1-C2 evidence	Governance constraints, approved plan and contracts, IaC handoff, validation, deployment summary, workflow diagram	C4: DR curveball
4	C3 evidence and changed requirement	Revised affected artifacts, reviews and approvals, ADR, updated IaC or paper design, updated cost and diagram	C5: Load test
5	Deployed endpoint or fallback plan	agent-output/freshconnect/05-load-test-results.md	C6: Documentation
6	All prior artifacts and observed state	agent-output/freshconnect/07-operations-runbook.md plus one additional as-built document	C7: Diagnostics
7	Platform evidence and docs	agent-output/freshconnect/07-diagnostics-quick-card.md	C8: Showcase
8	All prior artifacts, reviews, approvals, and evidence	Live presentation and Q&A	Wrap-up

Artifact Handoff Contract

• Check the required input and current review evidence before selecting the owning main agent.
• Save each artifact at the expected path and keep machine-readable sidecars with their Markdown artifact.
• Resolve required Challenger findings before recording approval.
• Keep validation evidence separate from deployment authorization.
• Name assumptions, failures, and verification limits clearly.
• Leave usable evidence for the next challenge, not a successful-looking chat response alone.

Prompt Recipe

Use this short recipe when a challenge page tells you to prompt an agent:

Goal: [artifact or decision]
Context: [FreshConnect facts + prior artifact path]
Decisions now: [2-4 key trade-offs]
Return: [file name, format, and evidence required]

Mermaid & ADR Conventions

• Mermaid diagrams should show the services, data flow, boundaries, and failure or handoff path that matters for the challenge.
• Workflow diagrams should show decisions and feedback loops, not just a straight line.
• ADRs should always cover Context, Decision, Consequences, and Alternatives.
• Keep diagrams readable enough to present in Challenge 8 without rework.

Paper Exercise Rules

• Use the paper path only when deployment is genuinely blocked.
• Keep the same output artifact names where possible.
• Record the blocker, intended change, and expected outcome.
• Do not present design intent as deployed evidence.

---

Pro Tips

Challenge 3 — Mermaid Flowcharts:

````mermaid
graph TD
    A[Start] --> B[Decision]
    B -->|Yes| C[Deploy]
    B -->|No| D[Refine]
\```
````

Challenge 4 — ADR Template:

# ADR: Multi-Region Disaster Recovery

## Context

[Why DR is needed now]

## Decision

[What approach we chose]

## Consequences

[Cost, complexity, benefits]

Challenge 6 — Documentation Prompts:

• “Generate ops runbook for [audience]”
• “Create cost breakdown with optimization tips”
• “Document DR procedures for compliance audit”

Challenge 7 — Diagnostic Queries:

# Quick health check
az webapp show --name <app-name> --resource-group <rg> --query state

# Application Insights query
az monitor app-insights query --app <app-id> \
  --analytics-query "requests | where timestamp > ago(1h) | summarize avg(duration)"

---

Naming Conventions

Resource	Pattern	Max
Key Vault	kv-{project}-{env}-{suffix}	24
Storage	st{project}{env}{suffix}	24
SQL Server	sql-{project}-{env}-{suffix}	63
App Service	app-{project}-{env}-{region}	60

Generate unique suffix:

var uniqueSuffix = uniqueString(resourceGroup().id)

---

Security Checklist

• supportsHttpsTrafficOnly: true
• minimumTlsVersion: 'TLS1_2'
• allowBlobPublicAccess: false
• azureADOnlyAuthentication: true
• Managed Identity (no connection strings)

---

Budget Guide

Phase	Budget	Region
Challenges 1-3	€500/month	swedencentral
After Curveball	€700/month	+ germanywestcentral

---

Load Test Targets

Metric	Target
Concurrent Users	500
P95 Response	≤ 2 seconds
Error Rate	≤ 1%

---

Help

Problem	Solution
Agent not responding	Reload VS Code window
Bicep won’t compile	Check bicep build output
Name too long	Use take() function
Zone redundancy error	Use P1v3+ SKU

---

Team: ****_**** Score: _/130