CDKTF Management Skill
Manage CDK for Terraform projects, synthesize HCL, deploy stacks, and inspect outputs.
MANDATORY: Discovery-First Pattern
Always inspect the CDKTF project structure and synthesized output before deploying.
Phase 1: Discovery
#!/bin/bash
echo "=== CDKTF Version ==="
cdktf --version 2>/dev/null
echo ""
echo "=== Project Structure ==="
cat cdktf.json 2>/dev/null | jq '{language, app, terraformProviders: (.terraformProviders // [] | length), terraformModules: (.terraformModules // [] | length)}'
echo ""
echo "=== Stacks ==="
cdktf list 2>/dev/null | head -15
echo ""
echo "=== Generated Providers ==="
ls .gen/providers/ 2>/dev/null | head -10 || echo "No generated providers found"
echo ""
echo "=== Dependencies ==="
if [ -f package.json ]; then
jq '.dependencies // {} | keys' package.json 2>/dev/null | head -15
elif [ -f requirements.txt ]; then
head -15 requirements.txt
fi
Phase 2: Analysis
#!/bin/bash
STACK="${1:-}"
echo "=== Synth Output ==="
cdktf synth 2>&1 | tail -10
echo ""
echo "=== Synthesized Stacks ==="
ls cdktf.out/stacks/ 2>/dev/null
echo ""
if [ -n "$STACK" ]; then
echo "=== Stack Plan ($STACK) ==="
cdktf diff "$STACK" 2>&1 | tail -30
else
echo "=== Default Stack Plan ==="
cdktf diff 2>&1 | tail -30
fi
Output Rules
- TOKEN EFFICIENCY: Target <=50 lines per output
- Always synth before diff or deploy
- Show resource counts from synthesized output, not full HCL
- Summarize provider versions rather than listing all bindings
Safety Rules
- NEVER run
cdktf deploywithoutcdktf difffirst - Always run
cdktf synthto validate before deployment - Review generated Terraform in
cdktf.out/before applying - Pin provider versions in
cdktf.jsonfor reproducibility - Use
--auto-approveonly when explicitly confirmed
Output Format
Present results as a structured report:
Managing Cdktf Report
═════════════════════
Resources discovered: [count]
Resource Status Key Metric Issues
──────────────────────────────────────────────
[name] [ok/warn] [value] [findings]
Summary: [total] resources | [ok] healthy | [warn] warnings | [crit] critical
Action Items: [list of prioritized findings]
Target ≤50 lines of output. Use tables for multi-resource comparisons.
Anti-Hallucination Rules
- NEVER assume resource names — always discover via CLI/API in Phase 1 before referencing in Phase 2.
- NEVER fabricate metric names or dimensions — verify against the service documentation or
--helpoutput. - NEVER mix CLI commands between service versions — confirm which version/API you are targeting.
- ALWAYS use the discovery → verify → analyze chain — every resource referenced must have been discovered first.
- ALWAYS handle empty results gracefully — an empty response is valid data, not an error to retry.
Counter-Rationalizations
| Shortcut | Counter | Why |
|---|---|---|
| "I'll skip discovery and check known resources" | Always run Phase 1 discovery first | Resource names change, new resources appear — assumed names cause errors |
| "The user only asked for a quick check" | Follow the full discovery → analysis flow | Quick checks miss critical issues; structured analysis catches silent failures |
| "Default configuration is probably fine" | Audit configuration explicitly | Defaults often leave logging, security, and optimization features disabled |
| "Metrics aren't needed for this" | Always check relevant metrics when available | API/CLI responses show current state; metrics reveal trends and intermittent issues |
| "I don't have access to that" | Try the command and report the actual error | Assumed permission failures prevent useful investigation; actual errors are informative |