KALNET Agent Strategy: Visual Summary

The Complete Build Graph

Figure 1 — KALNET Complete Build Graph. Fourteen agents execute across four stages. Within each stage all agents run in parallel; a barrier gate at each stage boundary waits for every agent to complete before the next stage begins. The critical path determines the ~2-hour wall-clock time.
Trace ID Requirement Rationale Test Case
KALNET-BUILD-001 All agents within a stage shall execute in parallel Minimises total wall-clock time from 136 min (sequential) to ~46 min (parallel) TC-KALNET-BUILD-001
KALNET-BUILD-002 Stage transitions shall block until every agent in the current stage succeeds Prevents downstream stages from running on incomplete foundations TC-KALNET-BUILD-002
KALNET-BUILD-003 Agent 1.4 (n8n) shall degrade gracefully if 1.2 (Comms) is not yet available Allows parallel execution while honouring the soft dependency TC-KALNET-BUILD-003

Trace ID	Requirement	Rationale	Test Case
`KALNET-BUILD-001`	All agents within a stage shall execute in parallel	Minimises total wall-clock time from 136 min (sequential) to ~46 min (parallel)	`TC-KALNET-BUILD-001`
`KALNET-BUILD-002`	Stage transitions shall block until every agent in the current stage succeeds	Prevents downstream stages from running on incomplete foundations	`TC-KALNET-BUILD-002`
`KALNET-BUILD-003`	Agent 1.4 (n8n) shall degrade gracefully if 1.2 (Comms) is not yet available	Allows parallel execution while honouring the soft dependency	`TC-KALNET-BUILD-003`

Decision Tree: What Do You Build?

Figure 2 — Build Scope Decision Tree. Choose the deployment scope based on available time. Each option is self-contained: Stage 0 alone provides a functional foundation, while the full four-stage build delivers a production-ready system.

Agent Priority Matrix

Figure 3 — Agent Priority Matrix. Agents are plotted by impact (vertical) and complexity (horizontal). Foundation agents (0.1, 0.2) and CI/CD (1.3) cluster in the “do first” quadrant. Complex integration agents (Dashboard, Wizard, Tests) require careful planning. Lower-impact agents can be deferred.
Trace ID Requirement Rationale Test Case
KALNET-PRI-001 Foundation agents (0.1, 0.2) shall be completed before any Stage 1 agent begins Network and storage are hard prerequisites for all services TC-KALNET-PRI-001
KALNET-PRI-002 Agents in the “do later” quadrant may be deferred without blocking production readiness Avoids over-engineering on low-impact, high-complexity work TC-KALNET-PRI-002

Trace ID	Requirement	Rationale	Test Case
`KALNET-PRI-001`	Foundation agents (0.1, 0.2) shall be completed before any Stage 1 agent begins	Network and storage are hard prerequisites for all services	`TC-KALNET-PRI-001`
`KALNET-PRI-002`	Agents in the “do later” quadrant may be deferred without blocking production readiness	Avoids over-engineering on low-impact, high-complexity work	`TC-KALNET-PRI-002`

Dependency Chain (What Blocks What)

Figure 4 — Agent Dependency Chain. Solid arrows indicate blocking dependencies; dashed arrows indicate soft dependencies that can be stubbed. The critical path is: 0.1 Network → 0.2 Storage → 1.3 CI/CD → 1.5 Discovery → 2.1 Dashboard → 3.1 Tests → 3.2 Release.
Trace ID Requirement Rationale Test Case
KALNET-DEP-001 The build scheduler shall respect all blocking dependency edges before launching an agent Prevents agents from running against missing infrastructure TC-KALNET-DEP-001
KALNET-DEP-002 Soft dependencies shall allow stub implementations so that parallel execution is not blocked Maximises parallelism while preserving integration correctness TC-KALNET-DEP-002

Trace ID	Requirement	Rationale	Test Case
`KALNET-DEP-001`	The build scheduler shall respect all blocking dependency edges before launching an agent	Prevents agents from running against missing infrastructure	`TC-KALNET-DEP-001`
`KALNET-DEP-002`	Soft dependencies shall allow stub implementations so that parallel execution is not blocked	Maximises parallelism while preserving integration correctness	`TC-KALNET-DEP-002`

Time Budget Breakdown

Figure 5 — Parallel Time Budget. With full parallelism the build completes in approximately 46 minutes (10 + 12 + 12 + 12), compared to 136 minutes when agents run sequentially. Each bar represents the wall-clock duration of a single agent.
Trace ID Requirement Rationale Test Case
KALNET-TIME-001 The fully parallel build shall complete within 50 minutes of wall-clock time Fits within a single engineering session for iterative feedback TC-KALNET-TIME-001
KALNET-TIME-002 Critical-path agents shall be prioritised if resources are constrained Ensures the build progresses even under reduced parallelism TC-KALNET-TIME-002

Trace ID	Requirement	Rationale	Test Case
`KALNET-TIME-001`	The fully parallel build shall complete within 50 minutes of wall-clock time	Fits within a single engineering session for iterative feedback	`TC-KALNET-TIME-001`
`KALNET-TIME-002`	Critical-path agents shall be prioritised if resources are constrained	Ensures the build progresses even under reduced parallelism	`TC-KALNET-TIME-002`

Quick Option Selector

Figure 6 — Quick Option Selector. Five build strategies ranging from zero-effort MVP shipping to full production readiness. Choose based on available time and immediate goals.

Parallelization Breakdown

Can these run at the same time?

STAGE 0:
1 (Network)    ↔ 0.2 (Storage)    ✓ Yes (independent)
1 (Network)    ↔ 0.3 (Observ)     ✓ Yes (independent)
1 (Network)    ↔ 0.4 (Security)   ✓ Yes (independent)
2 (Storage)    ↔ 0.3 (Observ)     ✓ Yes (independent)
2 (Storage)    ↔ 0.4 (Security)   ✓ Yes (independent)
3 (Observ)     ↔ 0.4 (Security)   ✓ Yes (independent)

STAGE 1:
1 (Games)      ↔ 1.2 (Comms)      ✓ Yes (independent)
1 (Games)      ↔ 1.3 (CI/CD)      ✓ Yes (independent)
1 (Games)      ↔ 1.4 (n8n)        ✓ Yes (independent)
1 (Games)      ↔ 1.5 (Discovery)  ✓ Yes (independent)
2 (Comms)      ↔ 1.3 (CI/CD)      ✓ Yes (independent)
2 (Comms)      ↔ 1.5 (Discovery)  ✓ Yes (independent)
3 (CI/CD)      ↔ 1.4 (n8n)        ✓ Yes (independent)
3 (CI/CD)      ↔ 1.5 (Discovery)  ✓ Yes (independent)
4 (n8n)        ↔ 1.5 (Discovery)  ✓ Yes (independent)
4 (n8n)  uses  1.2 (Comms)        ⚠ Soft dep (can stub)

STAGE 2:
1 (Dashboard)  ↔ 2.2 (Wizard)      ✓ Yes (independent)
1 (Dashboard)  ↔ 2.3 (Docs)        ✓ Yes (independent)
2 (Wizard)     ↔ 2.3 (Docs)        ✓ Yes (independent)

STAGE 3:
1 (Tests)      ↔ 3.2 (Release)     ✓ Yes (independent)

CONCLUSION: All agents within a stage can run in parallel!

What Gets Built At Each Stage

After Stage 0 (Foundation)

✓ Traefik reverse proxy (centralized access)
✓ Enhanced docker-compose with health checks
✓ Storage architecture and initialization scripts
✓ Centralized logging and metrics collection
✓ Security foundation (secrets, auth, permissions)
✓ Environment configuration templates

YOU CAN: Deploy existing services + access via single URL
READY FOR: All Stage 1 services

After Stage 1 (Core Services)

✓ Everything from Stage 0, plus:
✓ Game streaming infrastructure (Sunshine)
✓ Encrypted family communications (Matrix)
✓ GitOps deployment pipeline
✓ n8n automation and workflow templates
✓ Enhanced discovery service with service mesh
✓ Documentation for all services

YOU CAN: Stream games, chat privately, deploy code, automate tasks
READY FOR: UI/UX polish and validation

After Stage 2 (Integration)

✓ Everything from Stages 0-1, plus:
✓ Unified web dashboard (service status, activity, metrics)
✓ Interactive setup wizard (first-time configuration)
✓ Complete user + developer documentation
✓ Quick reference guides

YOU CAN: Access everything from one dashboard, onboard new users easily
READY FOR: Testing and release procedures

After Stage 3 (Validation)

✓ Everything from Stages 0-2, plus:
✓ Integration test suite (automated validation)
✓ Performance benchmarks
✓ Release procedures (versioning, changelog)
✓ Deployment automation (zero-downtime)
✓ Rollback procedures (safety net)

YOU CAN: Deploy to production with confidence, auto-test on changes
READY FOR: Family deployment and expansion

Agent Success Metrics

Each agent should deliver:

Figure 7 — Agent Success Metrics. Every agent must deliver across four dimensions: code deliverables, quality standards, documentation, and integration readiness. These criteria form the exit-gate checklist for each agent.

How to Actually Run This

OPTION 1: Copy/Paste to Claude (Simplest)
├─ Open this executor guide
├─ Copy Agent 0.1 prompt from AGENT-STRATEGY.md
├─ Paste to Claude
├─ Wait for output
├─ Save files
├─ Repeat for 0.2, 0.3, 0.4
├─ Then stages 1, 2, 3

OPTION 2: Batch with timestamps
├─ Create batch script: agents-stage-0.txt
├─ Contains all 4 prompts for stage 0
├─ Run through agent executor in parallel
├─ Collect outputs

OPTION 3: Selective (As you need)
├─ Only build what you need now
├─ Add more agents as you hit limitations
├─ KALNET grows over time, not all at once

If You Do Nothing Else, Do This

MINIMUM VIABLE BUILD: ~30 mins

Just run STAGE 0:
  ✓ invoke_agent "0.1-networking"
  ✓ invoke_agent "0.2-storage"
  ✓ invoke_agent "0.3-observability"     (optional)
  ✓ invoke_agent "0.4-security"          (optional)

WHY: Stage 0 is the foundation. Everything else depends on it.
RESULT: Traefik, storage architecture, logging, and security are in place.

THEN: Deploy existing services on top (Jellyfin, n8n, Samba)
         + your new MVP from tonight

NEXT: Come back and run Stage 1 when you're ready to expand.

The Schedule (If You Have 2 Hours Tonight)

00:00 - 00:05  Read this document
00:05 - 00:30  Deploy MVP from earlier (docker-compose up)
                └─ Meanwhile, queue Stage 0 agents

00:30 - 01:00  Run STAGE 0 (4 agents in parallel)
                ├─ Traefik configured
                ├─ Storage architecture ready
                ├─ Observability logging
                └─ Security foundation

01:00 - 01:10  Validate Stage 0 outputs
01:10 - 01:55  Run STAGE 1 (5 agents in parallel)
                ├─ Game streaming
                ├─ Encrypted comms
                ├─ CI/CD pipeline
                ├─ n8n workflows
                └─ Service mesh

01:55 - 02:00  Victory lap! 🎉

TOTAL: KALNET MVP + Foundation + Core Services
       Fully functional, ready for polish

Print this page. Bookmark it. Reference it when running agents. You’ve got this! 🚀

Overview

Network

Agents

Budget

Dashboard

​KALNET Agent Strategy: Visual Summary

​The Complete Build Graph

​Decision Tree: What Do You Build?

​Agent Priority Matrix

​Dependency Chain (What Blocks What)

​Time Budget Breakdown

​Quick Option Selector

​Parallelization Breakdown

​What Gets Built At Each Stage

​After Stage 0 (Foundation)

​After Stage 1 (Core Services)

​After Stage 2 (Integration)

​After Stage 3 (Validation)

​Agent Success Metrics

​How to Actually Run This

​If You Do Nothing Else, Do This

​The Schedule (If You Have 2 Hours Tonight)

KALNET Agent Strategy: Visual Summary

The Complete Build Graph

Decision Tree: What Do You Build?

Agent Priority Matrix

Dependency Chain (What Blocks What)

Time Budget Breakdown

Quick Option Selector

Parallelization Breakdown

What Gets Built At Each Stage

After Stage 0 (Foundation)

After Stage 1 (Core Services)

After Stage 2 (Integration)

After Stage 3 (Validation)

Agent Success Metrics

How to Actually Run This

If You Do Nothing Else, Do This

The Schedule (If You Have 2 Hours Tonight)