Storybloq/storybloq

What it solves

AI coding assistants are typically stateless, meaning they forget decisions, progress, and blockers between sessions. Developers often rely on fragmented notes or custom files to maintain context, leading to repeated mistakes and wasted setup time. Storybloq provides a standardized way to persist project state, tickets, and lessons learned across sessions so AI agents can compound their work rather than starting from zero.

How it works

Storybloq implements a file-based convention where a .story/ directory stores project metadata in JSON and Markdown files. It integrates with AI clients like Claude Code and Codex through:

• CLI & MCP Server: Provides structured tools for AI agents to read and mutate project state without spawning subprocesses.
• Skills: Custom commands (e.g., /story or $story) that load project status, handovers, and open tickets at the start of a session.
• Autonomous Mode: A state machine that drives tickets through a lifecycle of planning, implementation, testing, and review.
• Usage-Limit Auto-Resume: Detects when a session is killed by AI provider usage limits and automatically resumes the work once the limit resets.
• Federation: Allows an orchestrator project to coordinate work across multiple repositories, tracking dependencies and cross-repo blockers.

Who it’s for

Developers using AI coding assistants (specifically Claude Code and Codex) who want to maintain long-term project memory and coordinate complex, multi-step tasks across multiple sessions or repositories.

Highlights

• Cross-session persistence: Uses a .story/ directory to track tickets, issues, roadmaps, and handovers.
• Autonomous workflows: Automates the full ticket lifecycle from plan to commit.
• Usage-limit recovery: Automatically resumes autonomous sessions after AI provider rate limits expire.
• Multi-repo coordination: Federation features for managing dependencies and blockers across different repositories.
• Local Coordination Bus: An optional protocol for an implementer agent and a reviewer agent to exchange findings.
• MCP Integration: Native Model Context Protocol support for seamless tool use by AI clients.