TURING — Autonomous State Machine for Cognitive Continuity — Claude Code Plugin

TURING v1.1.0

Autonomous State Machine for Cognitive Continuity

What's New

Open Threads Tracking

TURING now captures and restores incomplete tasks across sessions. Open threads are displayed at session restore, ensuring nothing falls through the cracks during long development cycles.

How it works:

• Tracks uncompleted tasks in .claude/sessions/context.md
• Displays open threads on session restore
• Mark completed with - [x] syntax

Changes

Type	Description
Feature	Open threads tracking for task continuity across sessions
Docs	Updated comparison table and token estimates

Upgrading

# Update from marketplace
/plugin update turing@agenisea-ai

No breaking changes. Existing state files are fully compatible.

Full Changelog

• 62833d6 feat: add open threads tracking and bump to v1.1.0

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

TURING — Autonomous State Machine for Cognitive Continuity — Claude Code Plugin

TURING v1.0.0

Autonomous State Machine for Cognitive Continuity

The first stable release of TURING — a Claude Code plugin that preserves session state across context compaction events, enabling cognitive continuity for long-running AI-assisted development sessions.

Highlights

• 🔄 Zero-config operation — Install and forget. TURING works automatically via Claude Code hooks.
• 📦 No external dependencies — Uses only bash, Python 3, and git (pre-installed on macOS/Linux).
• ⚡ Lightweight footprint — ~1250 tokens per restore, priority-filtered for efficiency.

Features

Feature	Description
Priority-Based Restore	CRITICAL/HIGH/MEDIUM/LOW/ARCHIVE levels for token efficiency
Token Budget Tracking	Estimates context consumption, warns when state is bloated
Auto Decision Extraction	Extracts decisions from transcript ("decided to...", "going with...")
State Archiving	Previous states archived before overwrite
TTY-Based Session Discovery	Multiple terminals maintain independent state
ADR Recording	Architecture Decision Records with TL;DR summaries
YAML Frontmatter	Machine-parseable metadata with MD5 checksums

Installation

# Add the marketplace
/plugin marketplace add agenisea/turing

# Install the TURING plugin
/plugin install turing@agenisea-ai

Commands

Command	Description
/turing-save	Manually preserve state and record ADRs
/turing-status	View memory status (sessions, tokens, ADRs)
/turing-status --global	View memory across all projects on workstation

Requirements

• Claude Code 1.0+
• Python 3 (pre-installed on macOS/Linux)
• Git (for state capture)

Documentation

See TURING.md for comprehensive documentation including conceptual foundation, architecture, hook system details, and troubleshooting.