shepherd-context

Manage project state and synchronization with a SQLite-backed registry.

Updated May 12, 2026
One-click install
npx skills add https://github.com/FL03/claude-shepherd --skill shepherd-context
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: shepherd-context
Source: https://github.com/FL03/claude-shepherd/tree/main/skills/context
Command: npx skills add https://github.com/FL03/claude-shepherd --skill shepherd-context

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires sqlite3, jq, gh, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill solves the problem of state drift and context loss in long-running engineering sessions by providing a centralized, SQLite-backed registry for project symbols, GitHub state, and operational locks.

Core Features & Use Cases

  • Project Registry: Maintains a canonical SQLite database for project-specific state, including issues, PRs, and memory entries.
  • Drift Prevention: Implements automated gates for duplication detection and sprint-wide audit findings.
  • Use Case: Use this Skill to synchronize project context across multiple subagents, ensuring that every agent operates on the same source of truth for issue status and architectural decisions.

Quick Start

Initialize the project context registry and database by running the shepherd context initialization command in your project root.

Frequently Asked Questions about shepherd-context

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent state drift across parallel engineering agents?

To prevent state drift across parallel agents, use a centralized SQLite-backed registry that tracks GitHub state and operational locks, ensuring every agent operates on the same source of truth for issue status and architectural decisions.

What is the best way to manage project context synchronization for long-running sprints?

Project context synchronization for long-running sprints is managed through a SQLite-backed registry that implements automated duplication detection gates and tracks session-specific locks, maintaining canonical project state and consistency across concurrent workflows.

Does this project state tracking approach require sqlite3 and git to function?

Yes, this project state tracking approach requires sqlite3 and git to maintain the canonical project context database, alongside jq and gh dependencies, ensuring proper state synchronization and registry initialization in your project root.

How do I initialize a project context registry for sprint pipelines?

To initialize a project context registry for sprint pipelines, run the shepherd context initialization command in your project root to set up the SQLite database for tracking issues, PRs, and memory entries.

Can I use this registry to track GitHub issues and PRs across multiple subagents?

Yes, you can use this registry to track GitHub issues, PRs, and memory entries across multiple subagents by maintaining a canonical SQLite database that synchronizes project context and ensures consistency across parallel agent lanes.

What happens when duplicate sprint audit findings are detected in project state?

When duplicate sprint audit findings are detected in project state, automated drift prevention gates trigger to flag the duplication, maintaining the integrity of the canonical SQLite-backed registry and preventing context loss in engineering sessions.