v2.0.0 - Vision Alignment GA

Vision Alignment GA - General-Purpose Enterprise Orchestrator

Major release transforming Bob's Brain from ADK-focused devops assistant into a general-purpose enterprise orchestration system.

Highlights

• Canonical Agent IDs with backwards-compatible aliases
• Enterprise Controls (R0-R4 risk tiers, policy gates, evidence bundles)
• Mission Spec v1 (declarative workflow-as-code)
• 103 new unit tests (303 total)

New Capabilities

Agent Identity System (Phase D)

• Canonical kebab-case IDs: bob, iam-orchestrator, iam-compliance, etc.
• Alias support for legacy underscore IDs
• Deprecation warnings for migration path

Enterprise Controls (Phase E)

• Risk Tiers R0-R4: From read-only (R0) to financial operations (R4)
• Policy Gates: Default-deny for high-risk operations
• Tool Allowlists: Per-mandate tool restrictions
• Evidence Bundles: Complete audit trails with hashing

Mission Spec v1 (Phase F)

• Declarative YAML workflow definitions
• Deterministic compilation (same input → same output)
• CLI: validate, compile, dry-run, run
• Loop constructs with gates

New Documentation

• 252-DR-STND-agent-identity-standard.md
• 253-DR-STND-mandates-budgets-approvals.md
• 254-DR-STND-policy-gates-risk-tiers.md
• 255-DR-STND-evidence-bundles-and-audit-export.md
• 257-DR-STND-mission-spec-v1.md
• 260-AA-REPT-vision-alignment-ga-aar.md

Breaking Changes

None. Full backwards compatibility via alias system.

---

Full Changelog: See CHANGELOG.md

🤖 Generated with Claude Code

v1.1.0 - Google Multi-Agent Patterns Rollout

🎉 v1.1.0 - Google Multi-Agent Patterns Rollout

This release implements all 5 phases of the Google Multi-Agent Patterns rollout, bringing Bob's Brain from 3/8 to 7/8 implemented patterns (87.5%) from Google's Developer Blog guide.

✨ Features

Phase P1: Sequential Workflow Pattern

• ADK Primitive: SequentialAgent
• State passing via output_key and {key} reference syntax
• create_sequential_pipeline() factory function

Phase P2: Parallel Fan-Out Pattern

• ADK Primitive: ParallelAgent
• Unique output_key per agent to prevent collisions
• create_parallel_analyzer() factory function

Phase P3: Quality Gates Pattern

• ADK Primitive: LoopAgent
• Generator-Critic iteration with ctx.actions.escalate = True
• Configurable max_iterations
• create_quality_loop() factory function

Phase P4: Human-in-the-Loop Pattern

• ADK Primitive: Callbacks + before_agent_callback
• RiskLevel enum (LOW, MEDIUM, HIGH, CRITICAL)
• Approval request/response contracts
• Risk classification engine

Phase P5: Agent Starter Pack Templates

• 5 reusable pattern templates in templates/ directory
• Decision tree guide (PATTERNS.md)
• Contribution guidelines (CONTRIBUTING.md)

📝 Other Changes

• License: Changed from MIT to Elastic License 2.0 (ELv2)
• CI/CD: Fixed workflow references and ARV import checks

📊 Pattern Implementation Status

Pattern	Status
Sequential Pipeline	✅ Implemented
Parallel Fan-Out	✅ Implemented
Quality Gates	✅ Implemented
Human-in-the-Loop	✅ Implemented
Agent-to-Agent (A2A)	✅ Already existed
Foreman-Worker	✅ Already existed
Iterative Refinement	✅ Via LoopAgent
Routing/Delegation	❌ Not yet

📚 Documentation

• Added 6767-DR-STND-human-approval-pattern.md
• Added release report 236-AA-REPT-bobs-brain-release-v1-1-0.md
• Updated README badges (version + license)

Full Changelog: v1.0.0...v1.1.0

v0.13.0 - Linux Foundation AI Card PR Preparation

🎯 Linux Foundation AI Card PR Preparation

This release prepares Bob's Brain as a reference implementation for the Linux Foundation AI Card standard. All critical issues identified in the pre-PR audit have been resolved.

🚀 Highlights

• AI Card Reference Examples - 4 comprehensive files showing AI Card v1.0 adoption
• OSS Standard Files - CONTRIBUTING, CODE_OF_CONDUCT, SECURITY (25.4 KB total)
• Repository Quality - 85/100 → 95/100 (10-point improvement)
• R6 Compliance - Single docs folder achieved
• Documentation - All duplicates resolved, all links fixed

📦 What's New

AI Card Examples Directory

• ai-card-examples/bobs-brain/ with complete reference implementation
• Universal AI Card format (v1.0) with SPIFFE identity
• Original A2A AgentCard (v0.3.0) for comparison
• Comprehensive conversion guide
• Production patterns demonstrated (Hard Mode, Inline Deployment, Dual Memory)

OSS Standards

• CONTRIBUTING.md - Complete development guide (556 lines)
• CODE_OF_CONDUCT.md - Contributor Covenant v2.0
• SECURITY.md - Vulnerability reporting and security practices

Documentation Improvements

• Fixed 8 duplicate document numbers (history preserved)
• Updated 6767 master index with all 28 canonical standards
• Resolved all broken links
• Achieved R6 compliance (docs/ → github-pages/)

📊 Metrics

• Documentation: 145 files (141 → 145)
• Quality Score: 95/100 (85 → 95)
• Test Coverage: 65%+
• Hard Mode Compliance: 100%
• Broken Links: 0 (6 → 0)
• R6 Violations: 0 (1 → 0)

🔗 Key Documentation

• Linux Foundation PR Preparation AAR
• Repository Quality Audit
• 6767 Master Index
• AI Card Examples README

✅ Next Steps

• Submit PR to Linux Foundation AI Card repository
• Community review and feedback
• Potential adoption as canonical reference implementation

Full Changelog: v0.12.0...v0.13.0

v0.11.0: Production-Ready A2A Protocol & Monitoring

🚀 Release v0.11.0

Milestone: Agent Engine Deployment Readiness

This release represents full production readiness for Vertex AI Agent Engine deployment with comprehensive A2A protocol support and built-in monitoring strategy.

✨ Highlights

• Full A2A Protocol Implementation - Foreman AgentCard created with complete agent-to-agent discovery
• Built-in Monitoring Discovery - Vertex AI Agent Engine provides automatic metrics (no custom infrastructure needed!)
• Production CI/CD Pipeline - ARV gates, smoke tests, and deployment automation ready
• 171 Tests Passing - Comprehensive test coverage with AgentCard validation

📊 Key Metrics

• 59 commits since v0.10.0
• 2 major phases completed (21 & 22)
• 4 comprehensive AARs documenting all work
• 0 custom monitoring infrastructure needed (using built-in)

🔄 Changes

See CHANGELOG.md for detailed changes.

📦 What's Included

• Foreman agent ready for deployment
• Complete operator runbooks
• Monitoring strategy documentation
• Production-grade CI/CD workflows

🚀 Next Steps

1. Deploy agents to Vertex AI Agent Engine
2. Create monitoring dashboards using built-in metrics
3. Begin worker agent implementation

---

This is a canonical ADK reference implementation.

v0.9.0: Portfolio Orchestration & Org-Wide Storage

🚀 Release v0.9.0: Portfolio Orchestration & Org-Wide Storage

Major release adding centralized org-wide storage and multi-repo portfolio orchestration capabilities.

✨ Major Features

Org-Wide Knowledge Hub (LIVE1-GCS)

• Terraform infrastructure for centralized GCS storage (intent-org-knowledge-hub-{env})
• Python storage module with graceful error handling
• GCS writer for portfolio summaries + per-repo results
• 36 comprehensive tests (100% pass rate)
• Readiness check script for validation
• Complete documentation (architecture + AAR)

Multi-Repo Portfolio Orchestration (PORT1-3)

• Portfolio orchestrator for cross-repository SWE audits
• CLI with rich export (JSON/Markdown)
• Enhanced repo registry with tags and metadata
• GitHub Actions workflow for automated audits
• ARV integration with portfolio mode

IAM Department Templates

• Multi-agent framework for easy porting to other repos
• Comprehensive guides (porting, integration, operations)
• Reusable patterns for ADK-based departments

📊 Impact

• 20 commits since v0.8.0
• 226 files changed
• 36 new tests (100% pass)
• 20+ documents added to 000-docs/
• 100% backward compatible (all features opt-in)

🔧 Technical Details

GCS Storage Layout:

gs://{bucket}/portfolio/runs/{run_id}/summary.json
gs://{bucket}/portfolio/runs/{run_id}/per-repo/{repo_id}.json

Feature Flags (Opt-In):

• ORG_STORAGE_ENABLED (Terraform) - Create GCS bucket
• ORG_STORAGE_WRITE_ENABLED (Runtime) - Enable writes
• ORG_STORAGE_BUCKET (Runtime) - Bucket name

New Commands:

# Portfolio audits
python3 scripts/run_portfolio_swe.py
python3 scripts/run_portfolio_swe.py --repos bobs-brain,diagnosticpro
python3 scripts/run_portfolio_swe.py --output report.json --markdown report.md

# Org storage readiness
python3 scripts/check_org_storage_readiness.py
python3 scripts/check_org_storage_readiness.py --write-test

📚 Documentation

See the comprehensive documentation in 000-docs/:

• Org Storage Architecture
• LIVE1-GCS Implementation AAR
• Portfolio Orchestrator AAR
• IAM Department Templates

🔮 Future Work

• LIVE-BQ Phase: BigQuery integration for SQL analytics
• LIVE2 Phase: Dev-only RAG (Vertex AI Search) + Agent Engine wiring
• Multi-Repo Rollout: DiagnosticPro, PipelinePilot integration

---

Full Changelog: v0.8.0...v0.9.0

v0.8.0 - Agent Factory Structure

🏭 Agent Factory Transformation

Bob's Brain has been transformed from a single-agent repository to a production-grade agent factory with clean organization and zero clutter.

🎯 Major Changes

Repository Structure:

• agents/bob/ - Bob orchestrator (was my_agent/)
• templates/ - Reusable agent scaffolds
• scripts/ - Organized by purpose (ci/, deployment/, adk-docs-crawler/)
• archive/ - Consolidated from 99-Archive/

Import Paths (Breaking Change):

• Updated: my_agent → agents.bob (5 files)

Cleanup:

• ✅ Removed empty directories (adk-a2a/, tmp/, adk/)
• ✅ Removed redundant files (.env.sample)
• ✅ Archived legacy scripts

📊 Impact

• Files Changed: 2,304 files (2,273 archive consolidation, 31 structure)
• Commits: 11 focused commits squashed to main
• Status: CTO-ready, production-grade

🚀 Ready For

The repository is now structured to support:

• agents/iam-adk/ - ADK specialist agent
• Future iam-* agent team
• Multi-agent orchestration

📚 Documentation

• Complete transformation plan: 000-docs/077-AA-PLAN-agent-factory-structure-cleanup.md
• Updated: CLAUDE.md, README.md, .gitignore

Full Changelog: https://github.com/jeremylongshore/bobs-brain/blob/main/CHANGELOG.md#080---2025-11-19

Archive: feature/vertex-ai-genkit-migration

Archived branch feature/vertex-ai-genkit-migration for historical reference. Original branch deleted to maintain repository hygiene. Code remains accessible via this tag and release.

Archive: feat/phase-2-a2a-cloudrun-bootstrap

Archived branch feat/phase-2-a2a-cloudrun-bootstrap for historical reference. Original branch deleted to maintain repository hygiene. Code remains accessible via this tag and release.

Archive: feat/phase-1-scaffold

Archived branch feat/phase-1-scaffold for historical reference. Original branch deleted to maintain repository hygiene. Code remains accessible via this tag and release.

ci-green-20250920-1908: feat(ci): enforce 65% minimum code coverage

CI green: ed097d5