Cross-platform AI agents for Dynatrace demo workflows. Works on VS Code Copilot, Claude Code, Cursor, Windsurf, and OpenAI GPTs.
npx dt-copilot-agents install vscode # VS Code Copilot
npx dt-copilot-agents install claude-code . # Claude Code
npx dt-copilot-agents install cursor . # Cursor
npx dt-copilot-agents install # All platforms at onceVS Code Copilot — Open Chat (Ctrl+Alt+I) → click the mode selector → pick Dashboard Generator → type your prompt
Claude Code — Type /dashboard in the chat input → select the slash command → provide your prompt
Cursor / Windsurf — Just ask in chat:
SRE dashboard for Atlas Steel
CISO dashboard for Apex Bank
CEO dashboard for Summit Industries
CIO dashboard for meeting today — prospect at Helix Life Sciences
interview me for a VP Engineering dashboard
# POC Value Stories (Mode 5) — real tenant data required
MTTR improvement POC dashboard — Apex Bank, MCP connected
Log analytics value story — show DT logs vs ELK for a bank
Prove proactive monitoring value for a telco CIO
Business observability dashboard — e-commerce customer, real data
The agent researches the company, generates a 20-tile dashboard, and deploys it to your Dynatrace tenant — all automatically.
Zero prerequisites. The agent auto-installs
dtctlCLI and authenticates via browser SSO if not already set up.
Generates and deploys realistic persona-specific Dynatrace dashboards — in under 5 minutes.
| Mode | Trigger | How it works |
|---|---|---|
| Mode 1 — Demo (default) | Any simple prompt | Generates 20-tile dashboard with inline data record() DQL. Zero data ingestion needed. Instant. |
| Mode 2 — Live Tenant | "real data", "their Dynatrace", or MCP connected | Discovers real entities via MCP → builds dashboard with live timeseries, fetch logs, fetch spans queries |
| Mode 3 — Synthetic Ingest | "ingest", "persistent demo" | Generates realistic data → ingests via MINT/BizEvents API → dashboard queries live data |
| Mode 4 — Interview-First | "interview me", "help me figure out" | Agent asks 6 discovery questions → infers persona, story arc, and layout → builds the perfect dashboard |
| Mode 5 — POC Value Stories | "MTTR", "correlation", "logs value", "business observability", "proactive", "prove value" | Runs discovery DQL against real tenant → builds 20-tile dashboard framed as a value narrative for POC sign-off |
Meeting intelligence: Say "meeting today with [company]" to unlock story-arc selection, audience targeting, and CTA-focused layouts for your specific meeting goal.
Purpose-built dashboards that frame Dynatrace data as business value — designed for POC phases where you need to prove ROI and close.
| Value Story | Trigger phrases | Audience | What it proves |
|---|---|---|---|
| 1 — MTTR Improvement | "MTTR", "incident response", "resolution time" | SRE, IT Head, CIO | How DT cuts mean-time-to-resolve with automatic root cause |
| 2 — Correlation & Root Cause | "correlation", "root cause", "blast radius", "deployment impact" | CTO, App Ops, SRE | Davis AI automatically links events across the full stack |
| 3 — Business Observability | "business observability", "revenue impact", "tie IT to business" | CEO, CIO, CTO | Real-time bridge between IT health and business KPIs |
| 4 — Proactive Monitoring | "proactive", "shift left", "before users", "Davis value" | CIO, SRE, IT Head | Davis detects anomalies before users feel them |
| 5 — Log Analytics & Intelligence | "logs value", "log analytics", "logs vs ELK", "DQL logs", "log story" | App Ops, SRE, CTO, Platform Eng | From log chaos to instant DQL search, pattern detection, and log→trace correlation |
Example prompts:
MTTR improvement POC dashboard — Apex Bank, MCP connected
Log analytics value story for a bank — show how DT logs beats ELK
Prove proactive monitoring value for a telco CIO
Business observability dashboard — e-commerce, real data
Smart layout variants auto-select based on persona and meeting context:
| Variant | Persona / Occasion | Character |
|---|---|---|
| A — Panorama | CEO, CIO, Board presentations | Tall KPIs, wide charts — executive sweep |
| B — Standard | General / default | Balanced asymmetric grid |
| C — Dense | SRE, Ops, VP Engineering | Compact tiles, maximum data density |
| D — Hero | Renewal, PoV, first meeting | One tile dominates — the number that matters most |
| E — Trend-First | Growth story, degradation story | Wider time-series tiles, trend is the narrative |
Supported Personas:
| Executive | Technical / Operations |
|---|---|
| CEO — Business Impact | SRE — Reliability & SLOs |
| CIO — IT Strategy (default) | IT Head — Infrastructure |
| CTO — Technology & Innovation | Application Ops — APM |
| CISO — Security & Compliance | MLOps — AI/Model Operations |
| Platform Engineering — Developer Experience | |
| VP Engineering — Delivery & Quality |
Supported Industries: E-Commerce · Manufacturing · SaaS/Platform · Financial Services · Retail · Life Sciences · Telco
Features:
- Researches the company automatically via web (real plant names, product names, regions)
- Generates 20-tile dashboards with inline
data record()DQL (Mode 1 — zero ingestion) - Connects to live Dynatrace tenant for real metrics (Mode 2 — via MCP)
- Ingests synthetic data via MINT/BizEvents API for persistent demos (Mode 3)
- Guided discovery interview → builds the right dashboard (Mode 4)
- 5 POC Value Story dashboards — MTTR, Correlation, Business Observability, Proactive, Log Analytics (Mode 5)
- 5 smart layout variants that auto-select by persona and meeting context
- 6 story arcs: Risk, ROI, Visibility, Business Impact, Growth, Observability
- Deploys directly to Dynatrace tenant via
dtctlCLI - Validates all DQL queries before deploying (via MCP
verify_dql) - Layout is a flat
"layouts": { "1": {x,y,w,h} }structure — correctly enforced
| Platform | Format | Install | Auto-deploy via DTCTL |
|---|---|---|---|
| VS Code Copilot | .agent.md + skill |
npx dt-copilot-agents install vscode |
Yes |
| Claude Code | CLAUDE.md + .claude/commands/ |
npx dt-copilot-agents install claude-code |
Yes |
| Cursor | .cursor/rules/*.mdc |
npx dt-copilot-agents install cursor |
Yes |
| Windsurf | .windsurfrules |
npx dt-copilot-agents install windsurf |
Yes |
| M365 Copilot | Declarative Agent | Teams Toolkit deploy | Via Power Automate |
| OpenAI GPT | System prompt | Create Custom GPT | No (manual deploy) |
Auto-setup: The agent automatically installs
dtctland authenticates via browser SSO if not already set up — no manual prerequisite steps needed.
npx dt-copilot-agents install vscode # VS Code Copilot
npx dt-copilot-agents install claude-code . # Claude Code (current directory)
npx dt-copilot-agents install cursor . # Cursor
npx dt-copilot-agents install # All platforms at once
npx dt-copilot-agents info # Show supported personas & archetypesnpx dt-copilot-agents@latest install vscodegit clone https://github.com/pushpendrasinghbaghel-ai/dt-copilot-agents.git
cd dt-copilot-agents
.\install.ps1 # All platforms
.\install.ps1 -Platform vscode # VS Code onlyOpenAI Custom GPT — see openai/README.md
M365 Copilot — see m365-copilot/README.md
dt-copilot-agents/
├── knowledge/ # Shared knowledge (platform-agnostic)
│ └── dashboard-generator.md # Complete procedure, DQL rules, layout grid
│
├── agents/ # VS Code Copilot
│ └── dashboard-generator.agent.md
├── skills/ # VS Code Copilot skill
│ └── dt-demo-dashboard/
│ ├── SKILL.md
│ └── assets/
│
├── claude-code/ # Claude Code
│ ├── CLAUDE.md
│ └── .claude/commands/dashboard.md
│
├── cursor/ # Cursor
│ └── .cursor/rules/dashboard-generator.mdc
│
├── windsurf/ # Windsurf
│ └── .windsurfrules
│
├── m365-copilot/ # Microsoft 365 Copilot
│ ├── declarativeAgent.json
│ ├── instruction.md
│ └── README.md
│
├── openai/ # OpenAI / ChatGPT
│ ├── system-prompt.txt
│ └── README.md
│
├── install.ps1 # Cross-platform installer
└── README.md
All platforms share the same knowledge base (knowledge/dashboard-generator.md). Each platform gets a thin wrapper in its native format that references this shared knowledge. When you update the knowledge, all platforms benefit.
- Web access (for company research)
- Platform-specific: VS Code + Copilot extension, Claude Code CLI, Cursor IDE, or Windsurf IDE
macOS/Linux (Homebrew):
brew install dynatrace-oss/tap/dtctlmacOS/Linux (shell script):
curl -fsSL https://raw.githubusercontent.com/dynatrace-oss/dtctl/main/install.sh | shWindows (PowerShell):
irm https://raw.githubusercontent.com/dynatrace-oss/dtctl/main/install.ps1 | iexVerify: dtctl version
# Opens your browser for Dynatrace SSO — no tokens to copy-paste
dtctl auth login
# Or for a specific context
dtctl auth login --context <your-context-name>Alternative (CI/automation): dtctl auth login --token <API_TOKEN>
The agent uses your current DTCTL context by default. Switch contexts with:
dtctl config use-context <your-context-name>MCP enables the agent to verify DQL queries against your live tenant. It's optional — deployment works without it.
- Command Palette → MCP: Add Server → search "Dynatrace" → install
- Enter your tenant URL when prompted — authenticates via browser SSO
URL: https://<YOUR_TENANT>.apps.dynatrace.com/platform-reserved/mcp-gateway/v0.1/servers/dynatrace-mcp/mcp
Auth: Bearer <YOUR_PLATFORM_TOKEN>
Generate a platform token at: https://<YOUR_TENANT>.apps.dynatrace.com/ui/apps/dynatrace.classic.tokens
See knowledge/dashboard-generator.md → Authentication section for platform-specific config snippets (VS Code mcp.json, Claude Code .mcp.json, Cursor .cursor/mcp.json).
MIT — see LICENSE.