harvest

Package AI work sessions into structured handoff summaries with decisions, risks, and next steps.

Updated Feb 8, 2026
One-click install
npx skills add https://github.com/rbtkhn/grace-mar --skill harvest-rbtkhn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: harvest
Source: https://github.com/rbtkhn/grace-mar/tree/main/.cursor/skills/harvest
Command: npx skills add https://github.com/rbtkhn/grace-mar --skill harvest-rbtkhn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the challenge of transferring complex working context between AI sessions by creating a structured harvest packet that preserves decisions, artifacts, risks, and next steps without treating the output as durable memory.

Core Features & Use Cases

  • Session Context Extraction: Synthesizes visible conversation arcs, on-disk state, and relevant project signals into a compact handoff packet.
  • Agent Handoff Preparation: Creates structured markdown exports for parallel agents, review sessions, and downstream workflows while preserving uncertainty and scope boundaries.
  • Workflow Guardrails: Prevents accidental record mutation, unauthorized merges, and overclaiming transcript access during context transfer.

Quick Start

Use the harvest skill to create a session handoff packet summarizing the current work, decisions, risks, and recommended next steps.

Frequently Asked Questions about harvest

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

FAQPage Schema
What is a session handoff packet for AI agent collaboration?

A session handoff packet is a structured markdown summary that extracts decisions, artifacts, risks, and next steps from an ongoing AI work session to transfer context seamlessly between agents without treating the output as durable memory.

How do I transfer working context between parallel AI agents?

You can transfer working context by extracting visible conversation arcs, on-disk state, and project signals into a compact markdown handoff packet, which preserves uncertainty and scope boundaries for downstream agent collaboration.

Does context extraction for agent handoffs preserve git awareness?

Yes, context extraction for session handoffs applies git awareness to synthesize on-disk project state and relevant signals, ensuring the handoff packet accurately reflects the current repository conditions without mutating canonical records.

What are the limitations of using markdown packets for session continuity?

Markdown packets for session continuity are not durable memory and must avoid accidental record mutation, unauthorized merges, and overclaiming transcript access during context transfer between agents.

When do I need to package session insights for workflow automation?

You need to package session insights for workflow automation when performing parallel reviews, downstream agent collaboration, or continuity workflows that require disciplined context extraction and separation between session summaries and canonical records.

Can I use session handoff summaries for parallel review sessions?

Yes, session handoff summaries create structured markdown exports specifically designed for parallel agents and review sessions, preserving uncertainty and scope boundaries while transferring project files, decisions, artifacts, and risks.