a0.1.3

What's Changed

• 780dc1a docs: align documentation with gephyros architecture and ui
• 95d761a test: fix E0063 and unused variable warnings in test suite
• 9e3fdf2 refactor: overhaul system prompt to define core being, operational principles, and sub-agent delegation protocols
• bd3b4ab chore: update .gitignore to ignore the .agents directory instead of .agent
• 40f17cf Merge upstream-tracking into chore/align-system-prompts-with-cognition-2810462914
• 96b9afc fix: correct typos and table alignment in docs
• d69a51b feat(query): add coordinator_mode flag for orchestrator behavior
• 5d9c807 Refactor init command to use deep recon scan approach
• 47551e4 docs: Rename AGENT Framework to Gephyros Architecture in documentation
• f13b066 docs: add analysis of overrides, stop conditions, and behavioral drift
• 36f664a Merge branch 'chore/align-system-prompts-with-cognition-2810462914' into chore/docs-update-1037127493
• 6cc2b1e Added foundatonal lines to actions for better execution smoothness
• d0d464d docs: clarify Claurst Harness beta status and AGENT framework alpha release
• 48de476 Merge branch 'chore/docs-update-1037127493' into chore/align-system-prompts-with-cognition-2810462914
• accd536 refactor: refine system prompt capabilities and safety guidelines
• a276252 docs: Enhance agent self-discipline and confidence protocols
• 3483cf8 chore: add .clinerules/ to .gitignore
• f2378c1 docs: add detailed comparison of AGENT Framework vs standard prompt engineering
• 3cff7f7 feat(core): update system prompt with enhanced tool-use guidelines and security principle
• 5d0d404 refactor: Update system prompt for enhanced learning and project alignment
• 04a4ee6 docs(roadmap): add autonomic subconscious systems model design
• 4928fa7 docs: add cognition architecture documentation for memory, dreaming, and strange loop systems
• d2f68a8 docs(ATTRACTOR): add periodic nudge every 15 turns to file delivery modes
• 4c0f238 chore: update .gitignore to ignore .agent/ and .claurst/ directories
• 914bb29 docs: add COGNITION.md with system architecture diagrams
• 27e8fca fix: update repository URLs and document /init command
• ad514be docs: add five foundational seeds to BRAIN.md wisdom section
• 3b06dea docs: align project surface to Harness-Subsystem Architecture
• 949c915 docs: align docs with AGENT framework; deprecate CLAUDE.md
• 192cb1c docs: differentiate runtime agents from AGENT framework; add architecture docs
• 064a204 docs(readme): fix grammatical article in architecture description
• 4b63577 docs: update README to reference AGENT framework and refine Harness role
• 6e6e008 docs: add architecture documentation for harness-subsystem topology
• f4863f6 docs(attractor): expand with nudges, anatomy, and smooth gradient theory
• 69b4e27 docs: add semantic attractor architecture and update state pattern corpus
• a8597e0 feat: add AGENT.md as session initialization and personality file

Binaries

• claurst-linux-aarch64.tar.gz
• claurst-linux-x86_64.tar.gz

Installers

• install.sh (Linux/macOS)
• install.ps1 (Windows PowerShell)

v0.1.2

What's Changed

• 36f664a Merge branch 'chore/align-system-prompts-with-cognition-2810462914' into chore/docs-update-1037127493
• 6cc2b1e Added foundatonal lines to actions for better execution smoothness
• d0d464d docs: clarify Claurst Harness beta status and AGENT framework alpha release
• 48de476 Merge branch 'chore/docs-update-1037127493' into chore/align-system-prompts-with-cognition-2810462914
• accd536 refactor: refine system prompt capabilities and safety guidelines
• a276252 docs: Enhance agent self-discipline and confidence protocols
• 3483cf8 chore: add .clinerules/ to .gitignore
• f2378c1 docs: add detailed comparison of AGENT Framework vs standard prompt engineering

Binaries

• claurst-linux-aarch64.tar.gz (18MB)
• claurst-linux-x86_64.tar.gz (18MB)

Installers

• install.sh (Linux/macOS)
• install.ps1 (Windows PowerShell)

Version Scheme

• Claurst Harness (v0.1.4): Rust binary — TUI, CLI, providers, tools
• Claurst AGENT Framework (v0.1.2): System prompt, delivery timing, nudge, framework files

The Harness version is retained in Cargo.toml for upstream compatibility.

v0.1.1

What's Changed

• feat(core): update system prompt with enhanced tool-use guidelines and security principle
• refactor: Update system prompt for enhanced learning and project alignment

Binaries

• claurst-linux-aarch64.tar.gz (18M)
• claurst-linux-x86_64.tar.gz (18M)

v0.1 — AGENT Framework Integration

v0.1 — AGENT Framework Integration

This release marks the integration of the AGENT framework into the Claurst
harness, transforming it from a pure Claude Code port into a structured
cognitive architecture with explicit identity, memory, and attractor systems.

Architecture & Identity Layer:

• Added AGENT.md as session initialization and personality file (a8597e0)
• Added AGENTS.md, ATTRACTOR.md, BRAIN.md, HEART.md,
  STATE.md, USER.md — the full AGENT framework
  corpus defining the agent's identity, memory, and behavioral attractors
• Differentiated runtime named agents (--agent build|plan|explore) from
  the AGENT framework identity system (192cb1c)
• Aligned all documentation to Harness-Subsystem Architecture (3b06dea)
• Deprecated CLAUDE.md in favor of AGENT.md-based framework (949c915)

Semantic Attractor System:

• Added semantic attractor architecture with periodic nudge delivery
  every 15 turns (d2f68a8, 69b4e27, f4863f6)
• Hardcoded periodic_nudge_files to [AGENT.md, AGENTS.md, STATE.md,
  ATTRACTOR.md] for deterministic identity injection every turn

Memory & State:

• Added STATE.md as the project's durable, stable memory invariant map
• Added BRAIN.md with five foundational wisdom seeds (ad514be)

Infrastructure:

• Updated repository URLs from Kuberwastaken/claurst to
  acidgreenservers/claurst (27e8fca)
• Fixed install.sh and install.ps1 to point to new repo (27e8fca)
• Added .agent/workflows/build-binaries.md for reproducible builds
• Updated .gitignore to exclude .agent/ and .claurst/ directories (4c0f238)
• Built and verified claurst-linux-aarch64.tar.gz and
  claurst-linux-x86_64.tar.gz with all framework changes

This is the foundational release upon which the autonomic subconscious
systems (watcher agent, memory consolidation, dreaming loop) will be built.