Files
git.stella-ops.org/docs/implplan/SPRINT_5100_SUMMARY.md
master 53503cb407 Add reference architecture and testing strategy documentation
- Created a new document for the Stella Ops Reference Architecture outlining the system's topology, trust boundaries, artifact association, and interfaces.
- Developed a comprehensive Testing Strategy document detailing the importance of offline readiness, interoperability, determinism, and operational guardrails.
- Introduced a README for the Testing Strategy, summarizing processing details and key concepts implemented.
- Added guidance for AI agents and developers in the tests directory, including directory structure, test categories, key patterns, and rules for test development.
2025-12-22 07:59:30 +02:00

7.8 KiB

Sprint Epic 5100 · Comprehensive Testing Strategy

Overview

Epic 5100 implements the comprehensive testing strategy defined in the Testing Strategy advisory (20-Dec-2025). This epic transforms Stella Ops' testing moats into continuously verified guarantees through deterministic replay, offline compliance, interoperability contracts, and chaos resilience testing.

IMPLID: 5100 (Test Infrastructure) Total Sprints: 12 Total Tasks: ~75


Epic Structure

Phase 0: Harness & Corpus Foundation

Objective: Standardize test artifacts and expand the golden corpus.

Sprint Name Tasks Priority
5100.0001.0001 Run Manifest Schema 7 HIGH
5100.0001.0002 Evidence Index Schema 7 HIGH
5100.0001.0003 Offline Bundle Manifest 7 HIGH
5100.0001.0004 Golden Corpus Expansion 10 MEDIUM

Key Deliverables:

  • RunManifest schema capturing all replay inputs
  • EvidenceIndex schema linking verdict to evidence chain
  • BundleManifest for offline operation
  • 50+ golden test corpus cases

Phase 1: Determinism & Replay

Objective: Ensure byte-identical verdicts across time and machines.

Sprint Name Tasks Priority
5100.0002.0001 Canonicalization Utilities 7 HIGH
5100.0002.0002 Replay Runner Service 7 HIGH
5100.0002.0003 Delta-Verdict Generator 7 MEDIUM

Key Deliverables:

  • Canonical JSON serialization (RFC 8785 principles)
  • Stable ordering for all collections
  • Replay engine with frozen time/PRNG
  • Delta-verdict for diff-aware release gates
  • Property-based tests with FsCheck

Phase 2: Offline E2E & Interop

Objective: Prove air-gap compliance and tool interoperability.

Sprint Name Tasks Priority
5100.0003.0001 SBOM Interop Round-Trip 7 HIGH
5100.0003.0002 No-Egress Enforcement 6 HIGH

Key Deliverables:

  • Syft → cosign → Grype round-trip tests
  • CycloneDX 1.6 and SPDX 3.0.1 validation
  • 95%+ findings parity with consumer tools
  • Network-isolated test infrastructure
  • --network none CI enforcement

Phase 3: Unknowns Budgets CI Gates

Objective: Enforce unknowns-budget policy gates in CI/CD.

Sprint Name Tasks Priority
5100.0004.0001 Unknowns Budget CI Gates 6 HIGH

Key Deliverables:

  • stella budget check CLI command
  • CI workflow with environment-based budgets
  • PR comments with budget status
  • UI budget visualization
  • Attestation integration

Phase 4: Backpressure & Chaos

Objective: Validate router resilience under load.

Sprint Name Tasks Priority
5100.0005.0001 Router Chaos Suite 6 MEDIUM

Key Deliverables:

  • k6 load test harness
  • 429/503 response verification
  • Retry-After header compliance
  • Recovery within 30 seconds
  • Valkey failure injection tests

Phase 5: Audit Packs & Time-Travel

Objective: Enable sealed export/import for auditors.

Sprint Name Tasks Priority
5100.0006.0001 Audit Pack Export/Import 6 MEDIUM

Key Deliverables:

  • Sealed audit pack format
  • One-command replay verification
  • Signature verification with included trust roots
  • CLI commands for auditor workflow

Dependency Graph

Phase 0 (Foundation)
├── 5100.0001.0001 (Run Manifest)
│   └── Phase 1 depends
├── 5100.0001.0002 (Evidence Index)
│   └── Phase 2, 5 depend
├── 5100.0001.0003 (Offline Bundle)
│   └── Phase 2 depends
└── 5100.0001.0004 (Golden Corpus)
    └── All phases use

Phase 1 (Determinism)
├── 5100.0002.0001 (Canonicalization)
│   └── 5100.0002.0002, 5100.0002.0003 depend
├── 5100.0002.0002 (Replay Runner)
│   └── Phase 5 depends
└── 5100.0002.0003 (Delta-Verdict)

Phase 2 (Offline & Interop)
├── 5100.0003.0001 (SBOM Interop)
└── 5100.0003.0002 (No-Egress)

Phase 3 (Unknowns Gates)
└── 5100.0004.0001 (CI Gates)
    └── Depends on 4100.0001.0002

Phase 4 (Chaos)
└── 5100.0005.0001 (Router Chaos)

Phase 5 (Audit Packs)
└── 5100.0006.0001 (Export/Import)
    └── Depends on Phase 0, Phase 1

CI/CD Integration

New Workflows

Workflow Trigger Purpose
replay-verification.yml PR (scanner changes) Verify deterministic replay
interop-e2e.yml PR + Nightly SBOM interoperability
offline-e2e.yml PR + Nightly Air-gap compliance
unknowns-gate.yml PR + Push Budget enforcement
router-chaos.yml Nightly Resilience testing

Release Blocking Gates

A release candidate is blocked if any of these fail:

  1. Replay Verification: Zero non-deterministic diffs
  2. Interop Suite: 95%+ findings parity
  3. Offline E2E: All tests pass with no network
  4. Unknowns Budget: Within budget for prod environment
  5. Performance: No breach of p95/memory budgets

Success Criteria

Criteria Metric Gate
Full scan + attest + verify with no network offline-e2e passes Release
Re-running fixed input = identical verdict 0 byte diff Release
Grype from SBOM matches image scan 95%+ parity Release
Builds fail when unknowns > budget Exit code 2 PR
Router under burst emits correct Retry-After 100% compliance Nightly
Evidence index links complete Validation passes Release

Artifacts Standardized

Artifact Schema Location Purpose
Run Manifest StellaOps.Testing.Manifests Replay key
Evidence Index StellaOps.Evidence Verdict → evidence chain
Offline Bundle StellaOps.AirGap.Bundle Air-gap operation
Delta Verdict StellaOps.DeltaVerdict Diff-aware gates
Audit Pack StellaOps.AuditPack Compliance verification

Implementation Order

Immediate (This Week)

  1. 5100.0001.0001 - Run Manifest Schema
  2. 5100.0002.0001 - Canonicalization Utilities
  3. 5100.0004.0001 - Unknowns Budget CI Gates

Short Term (Next 2 Sprints)

  1. 5100.0001.0002 - Evidence Index Schema
  2. 5100.0002.0002 - Replay Runner Service
  3. 5100.0003.0001 - SBOM Interop Round-Trip

Medium Term (Following Sprints)

  1. 5100.0001.0003 - Offline Bundle Manifest
  2. 5100.0003.0002 - No-Egress Enforcement
  3. 5100.0002.0003 - Delta-Verdict Generator

Later

  1. 5100.0001.0004 - Golden Corpus Expansion
  2. 5100.0005.0001 - Router Chaos Suite
  3. 5100.0006.0001 - Audit Pack Export/Import


Execution Log

Date (UTC) Update Owner
2025-12-21 Epic created from Testing Strategy advisory analysis. 12 sprints defined across 6 phases. Agent

Epic Status: PLANNING (0/12 sprints complete)