Skip to content

AGET v2.9.0 - Information Storage Standardization

Choose a tag to compare

View all tags
@gmelli gmelli released this 24 Nov 06:25
· 170 commits to main since this release
v2.9.0
a81c2a6

AGET v2.9.0 - Information Storage Standardization

Overview

v2.9.0 introduces a formal 5-layer architecture for information storage, establishing clear boundaries between framework knowledge, agent types, instances, memory, and domain-specific content. This release includes comprehensive migration tools, formal specifications, and 100% fleet compliance across 28 agents.

Added

5-Layer Architecture

  • Layer 1 (Framework): Portable process knowledge in .aget/ directory
  • Layer 2 (Agent Type): Role-specific capabilities from templates
  • Layer 3 (Instance): Agent identity and configuration
  • Layer 4 (Memory): Advisory engagement state (.memory/ for advisors)
  • Layer 5 (Domain): Principal's work product at root level

Session Metadata Standard v1.0

  • Formal specification for session documentation
  • YAML frontmatter with quantitative metrics
  • Standardized fields: exchanges, tool_calls, objectives, blockers, patterns_discovered
  • Validation tooling included

Production Migration Tools (6 tools)

  • sessions_relocator.py v1.0.0 - Automated session relocation (.aget/sessions/ → sessions/)
  • memory_initializer.py v0.5.0 - Memory layer initialization for advisor agents
  • knowledge_classifier.py v0.5.0 - Framework vs domain knowledge classification
  • sessions_metadata_generator.py v1.0.0 - Automated YAML frontmatter generation
  • validate_fleet_standards.py v0.5.0 - Fleet compliance validation
  • audit_fleet_knowledge_architecture.py v1.0.0 - Architecture compliance audit

Comprehensive Documentation (175KB)

  • Information Storage Migration Guide v1.0 (96KB)
  • Migration Tools Reference v1.0 (98KB)
  • Session Metadata Standard v1.0
  • Directory Naming Specification v1.0

Changed

  • Session Location: Moved from .aget/sessions/ to sessions/ at root level
  • Knowledge Classification: Separated framework knowledge (.aget/evolution/) from domain knowledge (knowledge/ at root)
  • Directory Structure: Hierarchical organization with 5 distinct layers

Deprecated

  • .aget/sessions/ location (use sessions/ at root instead)
  • Mixed knowledge storage (separate framework vs domain knowledge)

Fixed

  • Issue #179: Working Directories - Clarified Layer 5 (Domain) for principal's work product
  • Issue #166: Memory Layer - Established Layer 4 architecture for advisor engagement state
  • Issue #164: Session Protocol - Formalized Session Metadata Standard v1.0
  • Issue #171: knowledge/ Directory - Separated Layer 1 (framework) from Layer 5 (domain) beliefs
  • Issue #101: Session Notes Protocol - Standardized session documentation with metadata

Metrics

  • Fleet Compliance: 100% (28/28 agents across 5 portfolios)
  • Quality: 9.90/10 average (13 rated gates)
  • Efficiency: 67% under budget (29.58h actual vs 89h estimate)
  • Perfect Gates: 8 consecutive (4.5, 6, 7, 8.5, 9, 10, 12, 12.5) 🏆
  • Zero Drift: 100% compliance maintained from Gate 8.5 → Gate 10
  • Tool Success Rate: 100% (279 sessions processed, 0 failures)

Migration Path

v2.8.0 → v2.9.0 is backward compatible (non-breaking architecture extension).

For existing agents:

  1. Use provided migration tools in sequence
  2. Validate with validate_fleet_standards.py
  3. See: Information Storage Migration Guide v1.0

For new agents:

  • All new agents must be created at v2.9.0 or higher
  • Use updated templates with 5-layer architecture

Resources

  • Migration Guide: .aget/docs/INFORMATION_STORAGE_MIGRATION_GUIDE_v1.0.md
  • Tools Reference: .aget/docs/MIGRATION_TOOLS_REFERENCE_v1.0.md
  • Session Metadata Spec: .aget/docs/SESSION_METADATA_STANDARD_v1.0.md
  • Directory Naming Spec: .aget/specs/DIRECTORY_NAMING_SPEC_v1.0.yaml

AGET Framework v2.9.0 - Configuration & Lifecycle Management for Human-AI Collaborative Coding

Assets 2
Loading