Bc Gov Multi App Analyst

BC Gov Multi-App Analyst

Repeatable orchestration pattern for structured analysis of 2+ interconnected BC Gov OCP applications.

Pre-Flight Checklist

Before starting any phase, verify:

• oc whoami succeeds on source cluster (Silver, Gold, or other)
• gh auth status confirms GitHub CLI authenticated
• Target namespace prefixes confirmed (e.g. <prefix1>, <prefix2>, <prefix3>)
• GitHub repositories identified for each app
• Output repo scaffolded (see Phase 0)
• ocp-migration-toolkit cloned and collect.sh executable
• container-network-analysis-toolkit cloned and collect.sh executable (for cross-namespace flows)

Phase 0 — Repository Scaffold

Set up the analysis repository using rl-project-template with the rl-agents-n-skills submodule.

# Clone template and initialise
gh repo create <org>/<analysis-repo> --template rloisell/rl-project-template --private
git clone git@github.com:<org>/<analysis-repo>.git
cd <analysis-repo>

# Add agents submodule
git submodule add git@github.com:rloisell/rl-agents-n-skills.git .agents

# Create output directory structure
mkdir -p AI/OCP-CLI-CAPTURES/{<ns1>,<ns2>,<ns3>}
mkdir -p AI/REPO-CAPTURES
mkdir -p docs/skill-proposals
mkdir -p report/apps/{<app1>,<app2>,<app3>}
mkdir -p report/cross-cutting
mkdir -p diagrams
mkdir -p scripts

Output structure:

<analysis-repo>/
├── AI/
│   ├── OCP-CLI-CAPTURES/
│   │   ├── <ns1>-dev/          # raw oc output per namespace
│   │   ├── <ns2>-dev/
│   │   └── <ns3>-dev/
│   ├── REPO-CAPTURES/          # GitHub repo snapshots + coupling matrix
│   └── nextSteps.md            # session continuity log
├── docs/
│   └── skill-proposals/        # Phase 4 skill extraction candidates
├── report/
│   ├── apps/
│   │   ├── <app1>/             # per-app deep-dive report
│   │   ├── <app2>/
│   │   └── <app3>/
│   └── cross-cutting/          # cross-app analysis reports
├── diagrams/                   # architecture + data flow diagrams
└── scripts/                    # collection scripts and helpers

Phase 1 — Parallel OCP + GitHub Capture

Run for each application namespace. Use ocp-migration-toolkit/collect/collect.sh for automated collection.

# Per namespace — automated collection
cd ocp-migration-toolkit
./collect/collect.sh \
  --namespace <prefix> \
  --cluster silver \
  --repo <org>/<repo> \
  --target emerald \
  --output ../AI/OCP-CLI-CAPTURES/

If manual collection is needed, run these commands per namespace:

NS=<prefix>-dev
oc get dc,deployment,statefulset,daemonset -n $NS -o yaml > AI/OCP-CLI-CAPTURES/$NS/workloads.yaml
oc get svc,route -n $NS -o yaml > AI/OCP-CLI-CAPTURES/$NS/network-svc-routes.yaml
oc get networkpolicy -n $NS -o yaml > AI/OCP-CLI-CAPTURES/$NS/networkpolicies.yaml
oc get pvc -n $NS -o yaml > AI/OCP-CLI-CAPTURES/$NS/pvcs.yaml
oc get secret -n $NS --no-headers | awk '{print $1, $2}' > AI/OCP-CLI-CAPTURES/$NS/secret-names.txt
oc get configmap -n $NS -o json > AI/OCP-CLI-CAPTURES/$NS/configmaps.json
oc get rolebinding -n $NS -o yaml > AI/OCP-CLI-CAPTURES/$NS/rolebindings.yaml
oc get pods -n $NS -o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{range .spec.containers[*]}{.image}{"\n"}{end}{end}' > AI/OCP-CLI-CAPTURES/$NS/image-refs.txt
oc get events -n $NS --sort-by='.lastTimestamp' | tail -30 > AI/OCP-CLI-CAPTURES/$NS/events.txt

For cross-namespace flow capture:

cd container-network-analysis-toolkit
./collect.sh --namespaces "<ns1>,<ns2>,<ns3>" --output ../AI/OCP-CLI-CAPTURES/

Phase 1b — Multi-App Coupling Matrix

See ocp-migration-analyst/SKILL.md Phase 1b for the full coupling matrix collection workflow.

Output file: AI/REPO-CAPTURES/coupling-matrix.md

The coupling matrix must be produced before Phase 2 per-app deep-dives and before Phase 3 cross-cutting reports, because:

• It determines migration sequencing (most upstream app first)
• It identifies shared infrastructure that must be present in the target before any app cutover
• It surfaces cross-namespace NPs that affect network security posture

Phase 2 — Parallel Per-App Deep-Dives

For each application, produce a deep-dive report at report/apps/<app>/<program-area>-<app>-deep-dive-v1.md.

Run per-app analyses in parallel where apps have no write-dependency on each other's output.

Each deep-dive report has 9 sections:

## 1. Executive Summary
## 2. Architecture Overview
## 3. Application Resiliency
## 4. Container Resiliency (R01–R15 against OCP Resilience Toolkit criteria)
## 5. Security
## 6. Network
## 7. BC Gov Standards Compliance
## 8. Findings Backlog (prioritised by risk)
## 9. Sources and Evidence

Per-App Sub-Agent Invocations

For each deep-dive, invoke these sub-agents with the collected data as context:

Phase 3 — Cross-Cutting Reports

Produce these reports at report/cross-cutting/:

Phase 4 — Skill/Toolkit Extraction

As patterns emerge during the analysis, document reusable knowledge for the skill library.

For each candidate pattern:

1. Create a file at docs/skill-proposals/<topic>.md
2. Document: the pattern observed, the concrete evidence, the generalised rule, and which skill to add it to
3. After engagement completion, raise PRs against rloisell/rl-agents-n-skills for each accepted skill update

Examples of reusable patterns extracted from past engagements:

• Cross-namespace NP risk classification → bc-gov-network-architect
• PII-in-logs checklist (Python/Go/Node.js) → observability
• RBAC CronJob anti-pattern → security-architect
• Image provenance risk tiers → security-architect
• External cloud data residency pattern → new data-flow-lineage skill

Phase 5 — Executive Summary and Packaging

Executive Summary

Produce report/<program-area>-executive-summary-v1.md with a RAG (Red/Amber/Green) traffic-light table:

| Dimension | <App1> | <App2> | <App3> |
|---|---|---|---|
| Application Resiliency | 🟡 | 🔴 | 🟡 |
| Container Resiliency | 🔴 | 🔴 | 🟡 |
| Security | 🔴 | 🔴 | 🟡 |
| Network | 🟡 | 🟡 | 🟢 |
| BC Gov Standards | 🟡 | 🟡 | 🟡 |
| Observability | 🟡 | 🟡 | 🟡 |
| Emerald Migration Readiness | 🔴 | 🔴 | 🟡 |

Additional Deliverables

• PROD-ACCESS-ASK.md — explicit list of production access required to complete findings that need prod data
• Diagrams: architecture overview, data flow, network topology (use diagram-generation agent)
• PDF render of all reports: pandoc <report>.md -o <report>.pdf --pdf-engine=wkhtmltopdf
• Git tag: git tag v1.0 && git push --tags

BC_GOV_MULTI_APP_ANALYST_KNOWLEDGE

• 2026-06-05: [multi-app-engagement] Parallel per-app deep-dives require coupling matrix first — sequencing without coupling matrix risks missing shared-infrastructure dependencies
• 2026-06-05: [multi-app-engagement] Phase 3 cross-cutting reports are higher value than per-app reports for executive stakeholders — prioritise these if time is constrained
• 2026-06-05: [multi-app-engagement] Skill extraction (Phase 4) is most effective when done continuously during the engagement, not as a post-engagement batch