dojo.md 0.1.0 → 0.2.1

package/courses/github-actions-cicd/scenarios/level-4/cicd-vendor-evaluation.yaml

@@ -0,0 +1,59 @@
+meta:
+  id: cicd-vendor-evaluation
+  level: 4
+  course: github-actions-cicd
+  type: output
+  description: "Evaluate CI/CD platforms — conduct a thorough comparison of GitHub Actions vs GitLab CI vs CircleCI vs Jenkins for enterprise adoption"
+  tags: [github, actions, vendor-evaluation, GitLab, CircleCI, Jenkins, expert]
+
+state: {}
+
+trigger: |
+  Your company is evaluating CI/CD platforms for a 500-engineer
+  organization. The current state: 60% use GitHub Actions, 25% use
+  Jenkins (legacy), and 15% use CircleCI (from an acquisition). The
+  CTO wants to consolidate to a single platform within 12 months.
+
+  Evaluation criteria (weighted by importance):
+  1. Feature completeness (25%): Workflow capabilities, integrations
+  2. Security (20%): OIDC, secret management, supply chain, audit
+  3. Scalability (15%): Concurrent jobs, runner management, performance
+  4. Cost (15%): Per-minute pricing, storage, data transfer
+  5. Developer experience (15%): YAML syntax, debugging, documentation
+  6. Enterprise features (10%): SSO, audit logs, compliance, governance
+
+  Platforms to evaluate:
+  A. GitHub Actions (current primary)
+  B. GitLab CI/CD (considered for integrated DevOps platform)
+  C. CircleCI (current secondary from acquisition)
+  D. Jenkins (current legacy, considering modern alternatives)
+
+  Scale requirements:
+  - 500 engineers, 200 repos
+  - 2,000 workflow runs per day
+  - Multi-cloud deployment (AWS, GCP, Azure)
+  - SOC 2 + PCI DSS compliance required
+  - Must support self-hosted runners for compliance
+  - Monorepo support needed for 3 teams
+
+  Task: Conduct the complete vendor evaluation. Write: the evaluation
+  matrix (score each platform on each criterion with justification),
+  the total cost of ownership analysis (3-year TCO including migration),
+  the risk assessment for each platform (vendor lock-in, outage history,
+  feature roadmap), the migration plan from the current 3-platform state
+  to the recommended single platform, and the executive recommendation
+  with decision rationale. Include the pilot program design.
+
+assertions:
+  - type: llm_judge
+    criteria: "Evaluation matrix is rigorous — each platform is scored on all 6 criteria with specific justifications (not generic), scores reflect real platform differences, and the weights are applied correctly to produce a weighted total"
+    weight: 0.35
+    description: "Rigorous evaluation matrix"
+  - type: llm_judge
+    criteria: "TCO analysis is complete — includes license/usage costs, migration effort (engineering hours), training costs, ongoing maintenance, and accounts for the different pricing models (per-minute vs per-user vs self-hosted). 3-year projection for all 4 options"
+    weight: 0.35
+    description: "Complete TCO analysis"
+  - type: llm_judge
+    criteria: "Executive recommendation is clear and defensible — makes a specific choice (not 'it depends'), addresses vendor lock-in risk, includes a migration timeline from 3 platforms to 1, and the pilot program validates the choice before full commitment"
+    weight: 0.30
+    description: "Clear executive recommendation"

package/courses/github-actions-cicd/scenarios/level-4/enterprise-cicd-governance.yaml

@@ -0,0 +1,55 @@
+meta:
+  id: enterprise-cicd-governance
+  level: 4
+  course: github-actions-cicd
+  type: output
+  description: "Design enterprise CI/CD governance — standardize workflows, enforce policies, and manage GitHub Actions across a large organization"
+  tags: [github, actions, enterprise, governance, policy, standardization, expert]
+
+state: {}
+
+trigger: |
+  You're the Director of Platform Engineering at a 500-engineer company
+  with 200+ repositories across 3 GitHub Enterprise organizations
+  (legacy from acquisitions). The CEO has mandated CI/CD standardization
+  after 3 incidents caused by inconsistent deployment practices.
+
+  Current chaos:
+  - 200+ repos with 400+ unique workflow files
+  - No standard deployment process (some repos deploy on merge, some
+    need manual approval, some have no CI at all)
+  - 15% of repos have no branch protection
+  - Third-party actions are used without vetting (supply chain risk)
+  - No visibility into CI/CD costs per team
+  - Different teams use different cloud providers with no standard auth
+  - Compliance requirements differ by BU (SOC 2, PCI, HIPAA)
+
+  Governance requirements:
+  1. Policy layer: Define what's allowed and required across all repos
+  2. Platform layer: Provide standard workflows teams can adopt
+  3. Enforcement layer: Automatically detect and remediate violations
+  4. Visibility layer: Dashboard showing compliance status per repo
+  5. Exception layer: Process for teams to request policy exceptions
+
+  Task: Design the enterprise CI/CD governance framework. Write: the
+  policy document (required controls for all repos, with BU-specific
+  extensions), the platform offerings (standard reusable workflows,
+  approved action allowlist, shared runner pools), the enforcement
+  automation (organization rulesets, required workflows, action
+  allowlist), the compliance dashboard design, and the exception
+  process. Include: the migration plan for the 200+ repos and the
+  organizational model (centralized platform team vs federated ownership).
+
+assertions:
+  - type: llm_judge
+    criteria: "Governance framework is comprehensive — covers policy definition, platform standardization, automated enforcement, visibility dashboards, and exception handling. The 5 layers work together cohesively"
+    weight: 0.35
+    description: "Comprehensive governance framework"
+  - type: llm_judge
+    criteria: "Enforcement is automated not manual — uses GitHub organization rulesets, required workflows, action allowlists (only approved actions can be used), and automated compliance scanning. Non-compliant repos are detected and flagged automatically"
+    weight: 0.35
+    description: "Automated enforcement"
+  - type: llm_judge
+    criteria: "Migration plan is realistic for 200+ repos — phased approach (not big-bang), provides standard workflows that teams can adopt incrementally, handles the 3-org consolidation, and includes success metrics"
+    weight: 0.30
+    description: "Realistic migration plan"

package/courses/github-actions-cicd/scenarios/level-4/expert-cicd-shift.yaml

@@ -0,0 +1,60 @@
+meta:
+  id: expert-cicd-shift
+  level: 4
+  course: github-actions-cicd
+  type: output
+  description: "Expert CI/CD shift — navigate vendor crises, compliance emergencies, and organizational challenges simultaneously"
+  tags: [github, actions, shift-simulation, crisis, vendor, compliance, expert]
+
+state: {}
+
+trigger: |
+  You're the VP of Platform Engineering during a week where 4 crises
+  converge on the CI/CD platform.
+
+  Crisis 1 — GitHub pricing change:
+  GitHub announced a 60% price increase for Actions minutes effective
+  in 90 days. Your current spend is $1.6M/year, projected to become
+  $2.56M/year. The CFO wants options: optimize, migrate to self-hosted,
+  or switch vendors. You need a cost mitigation plan by Friday.
+
+  Crisis 2 — Compliance audit finding:
+  The SOC 2 auditor discovered that 12% of production deployments have
+  no associated PR or code review (they were deployed via manual
+  workflow_dispatch triggers). This violates the change management
+  control. The auditor is asking for a corrective action plan within
+  2 weeks. The CTO wants to know how this happened.
+
+  Crisis 3 — Platform team attrition:
+  3 of your 15 platform engineers (including the runner fleet expert
+  and the reusable workflow architect) gave notice this week. They're
+  joining a competitor. Critical knowledge is at risk:
+  - Runner auto-scaling configuration (undocumented)
+  - Custom GitHub App for deployment orchestration (single maintainer)
+  - Vault integration for secret management (complex setup)
+
+  Crisis 4 — Developer revolt:
+  An internal blog post titled "Our CI/CD is Broken" got 200 upvotes.
+  Key complaints: "CI takes 30 minutes," "deployments fail randomly
+  and no one investigates," "the platform team says they'll fix it but
+  nothing changes." The CEO forwarded it asking for your response.
+
+  Task: Handle all 4 crises. For each, write: the immediate triage
+  (what to do today), the 2-week action plan, the 90-day resolution,
+  and the stakeholder communication. Then write the integrated strategy
+  memo showing how these crises are interconnected and the unified
+  approach to resolving them while maintaining team morale.
+
+assertions:
+  - type: llm_judge
+    criteria: "Cost mitigation plan is practical — analyzes options (optimize existing workflows, expand self-hosted, partial vendor migration), provides specific savings projections for each, and the recommendation balances cost with migration risk within the 90-day window"
+    weight: 0.35
+    description: "Practical cost mitigation"
+  - type: llm_judge
+    criteria: "Compliance remediation is achievable — addresses the workflow_dispatch bypass with branch protection and required approvals, provides the corrective action plan within the 2-week deadline, and prevents recurrence through automated enforcement"
+    weight: 0.35
+    description: "Achievable compliance remediation"
+  - type: llm_judge
+    criteria: "Integrated strategy connects all crises — shows how attrition impacts the other 3 crises (knowledge loss makes optimization harder), how the developer revolt is connected to underinvestment, and the unified approach addresses root causes rather than symptoms"
+    weight: 0.30
+    description: "Integrated crisis strategy"

package/courses/github-actions-cicd/scenarios/level-5/cicd-ai-future.yaml

@@ -0,0 +1,63 @@
+meta:
+  id: cicd-ai-future
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "Design the future of AI-powered CI/CD — architect intelligent pipelines that predict, prevent, and self-heal"
+  tags: [github, actions, AI, ML, intelligent-pipelines, future, master]
+
+state: {}
+
+trigger: |
+  You're the VP of AI/ML at a CI/CD platform company designing the
+  next generation of intelligent CI/CD. Your platform serves 5,000
+  organizations running 10M pipeline executions per month. You need to
+  design the AI system that defines CI/CD in 2028-2030.
+
+  Current AI capabilities (baseline):
+  - Predictive test selection: 60% test reduction, 95% fault detection
+  - Flaky test detection: Identifies flaky tests with 85% accuracy
+  - Basic failure classification: Categorizes failures (infra/code/dep)
+  - Cost recommendations: Suggests caching and parallelism improvements
+
+  Target capabilities (next 3 years):
+  1. Predictive CI: Before running CI, predict which tests will fail
+     based on the code diff, reducing CI from 20 min to 3 min
+  2. Self-healing pipelines: Automatically fix common failures (retry
+     transient errors, update cached dependencies, adjust timeouts)
+  3. Intelligent deployment: Predict deployment risk, automatically
+     choose deployment strategy (canary % based on risk), and auto-
+     rollback before users are impacted
+  4. Natural language pipelines: "Deploy to staging with extra load
+     testing" → generates and executes the workflow
+  5. Cross-organization learning: Learn patterns from anonymized data
+     across all customers to improve recommendations
+
+  Constraints:
+  - Customer pipeline code must never leave their tenant
+  - False positives in test selection must be < 1% (missing a real
+    failure is unacceptable)
+  - Natural language pipeline generation must be reviewed before
+    execution (no autonomous deployment without approval)
+  - Cross-org learning must use differential privacy
+
+  Task: Design the AI-powered CI/CD system. Write: the product vision
+  (what CI/CD looks like in 2030), the ML architecture (models,
+  training, inference, privacy), the safety framework (preventing AI
+  from causing deployments of broken code), the ethical considerations
+  (cross-org learning, developer surveillance), and the go-to-market
+  roadmap. Include the key technical challenges and proposed solutions.
+
+assertions:
+  - type: llm_judge
+    criteria: "Product vision is compelling and specific — describes concrete developer experiences in 2030 (not vague AI promises), shows the progression from current capabilities to target state, and acknowledges what AI cannot do (domain-specific testing, business logic validation)"
+    weight: 0.35
+    description: "Compelling specific vision"
+  - type: llm_judge
+    criteria: "ML architecture handles constraints — solves the privacy problem (federated learning or on-tenant models), achieves <1% false negative rate in test selection with confidence calibration, and the natural language pipeline generation includes human-in-the-loop review"
+    weight: 0.35
+    description: "Constraint-handling ML architecture"
+  - type: llm_judge
+    criteria: "Safety framework prevents AI-caused outages — defines clear boundaries (AI suggests, humans approve for production), includes circuit breakers for self-healing (rate limits on automated fixes), and addresses the ethical concerns of cross-org learning with differential privacy"
+    weight: 0.30
+    description: "Outage-preventing safety framework"

package/courses/github-actions-cicd/scenarios/level-5/cicd-behavioral-science.yaml

@@ -0,0 +1,70 @@
+meta:
+  id: cicd-behavioral-science
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "Apply behavioral science to CI/CD — use psychology and behavioral economics to improve developer adoption and reduce CI friction"
+  tags: [github, actions, behavioral-science, psychology, developer-experience, master]
+
+state: {}
+
+trigger: |
+  You're the Head of Developer Experience applying behavioral science
+  to CI/CD. Despite having excellent CI/CD infrastructure, developer
+  behavior isn't changing. You have the tools, but adoption and best
+  practices are lagging.
+
+  Observed behavioral patterns:
+  1. Learned helplessness: After years of slow, flaky CI, developers
+     don't trust the system even after improvements. 60% still run
+     tests locally before pushing (wasting 30 min/day) because "CI
+     might be broken again."
+
+  2. Present bias: Developers skip writing tests and security scans
+     to ship faster now, even though this causes 3x more incidents
+     later. The cost is delayed and distributed; the benefit is
+     immediate and personal.
+
+  3. Choice overload: The platform team offers 15 different workflow
+     templates, 8 runner types, and 12 optional integrations.
+     Developers choose the default (minimal) configuration 80% of
+     the time, missing valuable features.
+
+  4. Social proof gap: Teams with excellent CI/CD practices are
+     invisible. Teams don't know what "good" looks like because
+     there's no benchmarking or sharing of best practices.
+
+  5. Status quo bias: Teams resist migrating from Jenkins to GitHub
+     Actions (even when Actions is objectively better) because "Jenkins
+     works and we know it."
+
+  6. Feedback delay: When CI catches a bug, the feedback appears
+     15 minutes after the push. Developers have context-switched and
+     the feedback feels like an interruption rather than a helpful
+     catch.
+
+  Data available:
+  - 2 years of CI/CD usage data across 800 engineers
+  - Developer experience surveys (quarterly)
+  - A/B testing capability (workflow and UI modifications)
+
+  Task: Design a behavioral intervention program for CI/CD. For each
+  bias, write: the evidence from your data, the behavioral intervention
+  (nudge, default optimization, or environmental design), the
+  implementation in GitHub Actions (workflow changes, notifications,
+  gamification), and the A/B test design. Then write the unified
+  "Behavioral CI/CD" framework and the ethical considerations.
+
+assertions:
+  - type: llm_judge
+    criteria: "Interventions are behavior-specific — addresses learned helplessness with trust-building (CI reliability dashboard, streak counters), present bias with immediate feedback and CI-as-a-service defaults, choice overload with opinionated defaults and progressive disclosure"
+    weight: 0.35
+    description: "Behavior-specific interventions"
+  - type: llm_judge
+    criteria: "Implementation is practical in GitHub Actions — interventions are embedded in the CI/CD workflow itself (not separate tools), uses workflow summary annotations, PR comments, Slack notifications, and gamification elements that fit the developer workflow"
+    weight: 0.35
+    description: "Practical GitHub Actions implementation"
+  - type: llm_judge
+    criteria: "A/B test designs are rigorous — each has a clear hypothesis, control group, sample size consideration, success metric, and duration. Ethical considerations address developer privacy and the line between nudging and manipulation"
+    weight: 0.30
+    description: "Rigorous A/B test designs"

package/courses/github-actions-cicd/scenarios/level-5/cicd-board-strategy.yaml

@@ -0,0 +1,56 @@
+meta:
+  id: cicd-board-strategy
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "Board-level CI/CD strategy — present CI/CD as a strategic capability to the board of a public company"
+  tags: [github, actions, board, strategy, governance, competitive-advantage, master]
+
+state: {}
+
+trigger: |
+  You're the CTO of a $5B public technology company presenting to the
+  board of directors. The board has 3 CI/CD-related agenda items.
+
+  Agenda Item 1 — Competitive intelligence:
+  A competitor just published that they deploy 500 times per day with
+  99.99% success rate. Your company deploys 50 times per day with 95%
+  success. The board asks: "Are we falling behind? Is this a risk to
+  our market position?"
+  Your data: Your company releases features 40% faster year-over-year,
+  but the competitor's absolute velocity is 10x yours. However, your
+  change failure rate is improving while theirs has been static.
+
+  Agenda Item 2 — $15M infrastructure investment:
+  You're requesting $15M over 3 years to build an internal developer
+  platform with CI/CD at its core. The investment includes: platform
+  engineering team (20 FTEs), self-hosted runner infrastructure, custom
+  tooling, and AI-powered pipeline optimization.
+  The board asks: "What's the ROI? When does this pay back? What if
+  we just use GitHub Actions as-is without the platform investment?"
+
+  Agenda Item 3 — M&A technology assessment:
+  The company is acquiring a 500-engineer competitor. Due diligence
+  reveals: they use GitLab CI self-hosted, have no deployment
+  automation (manual deploys), and their CI takes 2 hours per run.
+  The board asks: "What's the integration risk? Timeline? Cost?"
+
+  Task: Prepare the board materials. For each agenda item, write: the
+  1-page board brief, the data visualizations (described in text), the
+  Q&A preparation (5 likely questions and answers), and the governance
+  recommendations. End with a unified strategic narrative connecting
+  all 3 items.
+
+assertions:
+  - type: llm_judge
+    criteria: "Competitive analysis is nuanced — doesn't panic about the 10x gap, explains that absolute deploy frequency is less important than rate of improvement and change failure rate, and connects CI/CD velocity to business outcomes (time-to-market, customer satisfaction)"
+    weight: 0.35
+    description: "Nuanced competitive analysis"
+  - type: llm_judge
+    criteria: "Investment case is board-ready — presents the $15M as a 3-year investment with clear payback timeline, quantifies the 'do nothing' risk (cost of not investing), and the ROI model includes both tangible returns (cost savings) and intangible (developer productivity, talent retention)"
+    weight: 0.35
+    description: "Board-ready investment case"
+  - type: llm_judge
+    criteria: "M&A assessment is realistic — quantifies the integration risk (timeline, cost, talent retention risk during migration), proposes a phased approach, and connects the platform investment (Item 2) to faster M&A integration capability"
+    weight: 0.30
+    description: "Realistic M&A assessment"

package/courses/github-actions-cicd/scenarios/level-5/cicd-consulting-engagement.yaml

@@ -0,0 +1,61 @@
+meta:
+  id: cicd-consulting-engagement
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "Lead a CI/CD consulting engagement — diagnose and transform a client's broken CI/CD infrastructure as an external consultant"
+  tags: [github, actions, consulting, transformation, engagement, master]
+
+state: {}
+
+trigger: |
+  You're a senior DevOps consultant hired for a 12-week, $200K
+  engagement to transform CI/CD at a 300-engineer healthcare company.
+  They're 18 months into a failed GitHub Actions migration from Jenkins
+  and are about to lose their largest customer ($40M contract) due to
+  compliance gaps.
+
+  Client diagnostic findings (Week 1):
+  - 200 repos: 80 still on Jenkins, 100 on GitHub Actions (inconsistent
+    config), 20 with no CI at all
+  - Average deployment: 2 per week (target: 10 per day)
+  - Failed deployments: 30% (target: < 5%)
+  - Mean time to recovery: 6 hours (target: < 30 minutes)
+  - No automated security scanning (HIPAA violation risk)
+  - Self-hosted runners run on unpatched VMs with shared credentials
+  - The Jenkins→Actions migration was led by one person who left
+  - No documentation of the current CI/CD architecture
+  - The compliance team manually generates audit reports (2 weeks/quarter)
+  - Developer satisfaction with CI/CD: 18% (industry benchmark: 65%)
+
+  Client stakeholders:
+  - CTO: "We need to pass the HIPAA audit in 90 days or we lose the
+    customer. Fix this."
+  - VP Engineering: "My teams are demoralized. They've been promised a
+    'better CI/CD' for 18 months and things got worse."
+  - CISO: "I have no visibility into what's deployed or how. Our
+    security posture is unacceptable."
+  - Engineering Manager: "The last consultant just wrote a document.
+    We need someone who will actually help us fix things."
+
+  Task: Design the complete consulting engagement. Write: the client
+  diagnostic report (executive summary with quantified risk), the
+  12-week transformation roadmap (phased with weekly deliverables),
+  the quick wins for Weeks 1-3 (building credibility and addressing
+  the HIPAA audit deadline), the technical implementation plan (complete
+  the Jenkins migration, standardize Actions, fix security), and the
+  handoff package (documentation, training, maintenance runbooks).
+
+assertions:
+  - type: llm_judge
+    criteria: "Diagnostic report quantifies risk — connects CI/CD failures to business impact ($40M customer at risk), identifies the root cause of the failed migration (single person, no documentation, no plan), and presents findings without blame"
+    weight: 0.35
+    description: "Risk-quantifying diagnostic"
+  - type: llm_judge
+    criteria: "Roadmap addresses the HIPAA deadline — Weeks 1-3 focus on the most critical compliance gaps (security scanning, audit trails, runner security), Weeks 4-8 complete the migration, Weeks 9-12 establish sustainability, and the 90-day HIPAA audit timeline is achievable"
+    weight: 0.35
+    description: "HIPAA-focused roadmap"
+  - type: llm_judge
+    criteria: "Handoff package ensures sustainability — includes documentation of all CI/CD architecture, runbooks for common operations, training curriculum for the team, and a monitoring dashboard so the client can maintain the system after the consultant leaves"
+    weight: 0.30
+    description: "Sustainability-ensuring handoff"

package/courses/github-actions-cicd/scenarios/level-5/cicd-industry-benchmarks.yaml

@@ -0,0 +1,63 @@
+meta:
+  id: cicd-industry-benchmarks
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "Publish CI/CD benchmarks — create the definitive 'State of CI/CD' report with cross-industry analysis and DORA insights"
+  tags: [github, actions, benchmarks, DORA, research, industry-analysis, master]
+
+state: {}
+
+trigger: |
+  You're the Head of Research at a developer tools company publishing
+  the annual "State of CI/CD 2026" report. This report is read by
+  60,000+ engineering leaders. You have data from 3,000 organizations.
+
+  Raw data highlights:
+  - DORA metrics by company size:
+    | Metric              | Startup | Mid-market | Enterprise |
+    |---------------------|---------|-----------|------------|
+    | Deploy frequency    | 12/day  | 3/day     | 2/week     |
+    | Lead time           | 1 hr    | 8 hrs     | 5 days     |
+    | MTTR                | 15 min  | 1 hr      | 8 hrs      |
+    | Change failure rate | 5%      | 8%        | 15%        |
+
+  - CI/CD platform market share:
+    GitHub Actions: 62% | GitLab CI: 18% | Jenkins: 12% |
+    CircleCI: 4% | Other: 4%
+
+  - Key correlations:
+    Teams with CI under 10 min: 3x more deploys, 50% lower failure rate
+    Teams using OIDC: 75% fewer credential incidents
+    Teams with automated rollback: 80% lower MTTR
+    Monorepo teams: 2x CI cost but 1.5x deployment frequency
+
+  - Emerging trends:
+    AI in CI/CD: 28% using AI for test selection, 15% for failure
+    prediction, 8% for automated remediation
+    Platform engineering: 45% have dedicated platform teams (up from 12%)
+    Supply chain security: 52% pin actions by SHA (up from 8%)
+
+  - Industry breakdown (deployment frequency):
+    SaaS: 15/day | E-commerce: 8/day | Fintech: 3/day |
+    Healthcare: 1/day | Government: 2/month | Gaming: 20/day
+
+  Task: Write the "State of CI/CD 2026" report. Include: the executive
+  summary, the benchmarking methodology, the DORA metrics analysis by
+  company size and industry, the emerging trends analysis (AI, platform
+  engineering, supply chain), the strategic recommendations, and the
+  predictions for 2027-2028.
+
+assertions:
+  - type: llm_judge
+    criteria: "Methodology is credible — describes sample selection, survey design, data validation, limitations (survivorship bias, self-reporting), and statistical significance. Acknowledges correlation ≠ causation for the key findings"
+    weight: 0.35
+    description: "Credible methodology"
+  - type: llm_judge
+    criteria: "Analysis reveals actionable insights — explains why enterprise DORA metrics lag (compliance overhead, organizational complexity), identifies the 10-minute CI threshold as a tipping point, and provides specific recommendations by company size"
+    weight: 0.35
+    description: "Actionable insight analysis"
+  - type: llm_judge
+    criteria: "Predictions are specific and bold — makes concrete predictions about AI adoption rates, platform engineering growth, GitHub Actions market share evolution, and the decline of Jenkins, with reasoning grounded in the trend data"
+    weight: 0.30
+    description: "Specific bold predictions"

package/courses/github-actions-cicd/scenarios/level-5/cicd-ma-integration.yaml

@@ -0,0 +1,73 @@
+meta:
+  id: cicd-ma-integration
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "M&A CI/CD integration — consolidate CI/CD platforms and practices across acquired companies"
+  tags: [github, actions, M&A, integration, consolidation, migration, master]
+
+state: {}
+
+trigger: |
+  You're the CTO overseeing CI/CD integration for 3 recently acquired
+  companies. Each uses a different CI/CD platform and has different
+  engineering practices. The board expects full integration within 18
+  months.
+
+  Parent company (1,000 engineers):
+  - GitHub Actions with sophisticated platform team
+  - DORA Elite: 50 deploys/day, 2hr lead time, 15min MTTR, 2% CFR
+  - Fully automated, OIDC auth, compliance automation
+  - Cost: $2M/year for CI/CD infrastructure
+
+  Acquisition A — SaaS startup (200 engineers):
+  - CircleCI with custom orbs
+  - Fast but fragile: 30 deploys/day but 15% failure rate
+  - No security scanning, no compliance controls
+  - 80 repos, all on CircleCI
+  - Cost: $400K/year
+
+  Acquisition B — Enterprise software (300 engineers):
+  - Jenkins on-premise (50+ plugins, custom Groovy pipelines)
+  - Slow but stable: 2 deploys/week, 3% failure rate
+  - Heavy compliance (SOX, HIPAA), manual audit processes
+  - 120 repos on Bitbucket + Jenkins
+  - Cost: $800K/year (hardware + maintenance)
+
+  Acquisition C — AI/ML company (150 engineers):
+  - Mix of GitHub Actions (for apps) and custom Python scripts (for
+    ML pipelines)
+  - ML model training uses separate GPU pipeline outside CI/CD
+  - Jupyter notebooks deployed via manual process
+  - 40 repos on GitHub + internal GitLab for ML experiments
+  - Cost: $300K/year
+
+  Integration constraints:
+  - Must consolidate to GitHub Actions within 18 months
+  - Cannot disrupt any acquisition's shipping velocity during migration
+  - Must achieve SOX + HIPAA compliance for all entities
+  - Must retain 90% of acquired engineering talent
+  - Budget: $3M for migration (separate from ongoing costs)
+  - ML pipeline integration is the most technically complex
+
+  Task: Design the M&A CI/CD integration strategy. Write: the assessment
+  of each acquisition's CI/CD maturity, the migration plan for each
+  (CircleCI → Actions, Jenkins → Actions, custom scripts → Actions),
+  the compliance harmonization plan (bringing all entities to SOX + HIPAA),
+  the ML pipeline integration strategy, and the talent retention plan.
+  Include: the 18-month timeline with quarterly milestones and the $3M
+  budget allocation.
+
+assertions:
+  - type: llm_judge
+    criteria: "Migration plans are tailored — CircleCI migration is relatively straightforward (similar paradigm), Jenkins migration requires complete rethinking (Groovy → YAML), ML pipeline migration needs custom approach (GPU runners, notebook deployment), and each has a specific timeline"
+    weight: 0.35
+    description: "Tailored migration plans"
+  - type: llm_judge
+    criteria: "Compliance harmonization is realistic — addresses the gap between startup's no-compliance and enterprise's heavy compliance, proposes a unified compliance framework that works for all entities, and the timeline achieves SOX + HIPAA compliance within the 18-month window"
+    weight: 0.35
+    description: "Realistic compliance harmonization"
+  - type: llm_judge
+    criteria: "Budget allocation and timeline are justified — the $3M is distributed across the 3 acquisitions proportionally to migration complexity, the timeline accounts for parallel workstreams, and the talent retention plan addresses flight risk during migration disruption"
+    weight: 0.30
+    description: "Justified budget and timeline"

package/courses/github-actions-cicd/scenarios/level-5/cicd-product-development.yaml

@@ -0,0 +1,68 @@
+meta:
+  id: cicd-product-development
+  level: 5
+  course: github-actions-cicd
+  type: output
+  description: "Build a CI/CD product — design and launch a CI/CD SaaS startup competing in the developer tools market"
+  tags: [github, actions, product, startup, SaaS, go-to-market, master]
+
+state: {}
+
+trigger: |
+  You're the co-founder/CTO of a startup building "PipelineHQ" — a
+  next-generation CI/CD platform. You've raised $8M in seed funding
+  and have 18 months of runway. Your thesis: GitHub Actions is powerful
+  but enterprise teams need better observability, cost management, and
+  governance on top of it.
+
+  Market analysis:
+  - TAM: $12B (CI/CD and DevOps tools market)
+  - SAM: $3B (CI/CD platforms and add-ons)
+  - SOM: $300M (GitHub Actions ecosystem tools)
+  - Competitors: Harness ($2B valuation), Buildkite ($200M), Earthly
+    ($20M), Depot ($5M), Trunk ($30M)
+  - Gap: No single tool combines Actions analytics + cost management +
+    governance in a developer-friendly package
+
+  Product vision:
+  1. Cost Intelligence: Real-time per-team cost tracking, optimization
+     recommendations, budget enforcement
+  2. Pipeline Analytics: DORA metrics, failure patterns, bottleneck
+     detection, workflow performance trending
+  3. Governance: Policy-as-code for Actions, action allowlisting,
+     compliance automation, audit reporting
+  4. Developer Experience: Smart caching, predictive failures, auto-
+     retry with root cause analysis
+
+  Technical constraints:
+  - Must work with GitHub Actions (not replace it)
+  - GitHub App integration (webhook-based, no code injection)
+  - Customer workflow data stays in their GitHub (privacy)
+  - Must handle 10,000+ workflow runs/day per customer
+
+  Go-to-market:
+  - 18-month runway ($8M, burn rate $450K/month)
+  - 10-person engineering team
+  - Need $2M ARR to raise Series A
+  - PLG + enterprise sales motion
+
+  Task: Design the complete product strategy. Write: the MVP roadmap
+  (3-month scope for first paying customer), the technical architecture
+  (GitHub App, webhook processing, analytics engine), the pricing
+  strategy (free/team/enterprise tiers), the GTM plan (PLG adoption
+  loop, enterprise sales playbook), and the competitive positioning.
+  Include the financial projection and Series A pitch narrative.
+
+assertions:
+  - type: llm_judge
+    criteria: "MVP scope is achievable in 3 months by 10 engineers — focuses on one killer feature (likely cost intelligence or analytics, not all 4), has a clear first customer profile, and the scope decision is justified with competitive analysis"
+    weight: 0.35
+    description: "Achievable MVP scope"
+  - type: llm_judge
+    criteria: "Technical architecture is sound — GitHub App with webhook processing handles the scale, respects data privacy constraints, the analytics engine can process 10K+ runs/day, and the architecture supports the product roadmap beyond MVP"
+    weight: 0.35
+    description: "Sound technical architecture"
+  - type: llm_judge
+    criteria: "Financial projection reaches $2M ARR in 18 months — PLG funnel conversion assumptions are realistic, pricing tiers incentivize upgrade, enterprise sales pipeline is modeled, and the burn rate / runway math is consistent"
+    weight: 0.30
+    description: "Realistic financial projection"