salvage

Extract structured learnings and restart context from drifting sessions.

3|Updated Jan 27, 2026
Ready to run
npx skills add https://github.com/open-horizon-labs/skills --skill salvage
💬 Prompt for ChatGPT / Claude Chat▼
Please help me install this Agent Skill.
Skill: salvage
Source: https://github.com/open-horizon-labs/skills/tree/main/salvage
Command: npx skills add https://github.com/open-horizon-labs/skills --skill salvage

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Salvage helps you extract learning from a session or piece of work before restarting, ensuring you don't lose insight when drift happens.

Core Features & Use Cases

  • Structured salvage workflow to name drift state, extract five learning categories, and package a restart kit.
  • Persist learnings to a session store or graph (OH MCP) when available for cross-skill reuse.
  • Generate restart-ready context: guardrails, missing context, local practices, and reusable fragments.

Quick Start

Describe the current drift and run salvage to extract actionable learnings.

Frequently Asked Questions about salvage

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

FAQPage Schema
How do I extract learning from a drifting session before restarting?▼

To extract learning from a drifting session, describe the current drift state to trigger a structured salvage workflow that packages learnings, guardrails, missing context, and reusable fragments into a restart-ready context.

What is the best way to salvage scope creep when progress stalls?▼

Salvaging scope creep when progress stalls requires naming the drift state and extracting five learning categories to generate a structured report with missing context, local practices, and restart context.

Why does work reverse multiple times and how can I save the insights?▼

Work reverses multiple times due to scope expansion and stalling progress, but you can save insights by applying a structured salvage workflow to extract learnings and generate a restart kit.

How do I persist extracted learnings for cross-skill reuse?▼

To persist extracted learnings for cross-skill reuse, the salvage workflow saves the generated guardrails and reusable fragments to a session store or graph when available.

What is included in a structured salvage report?▼

A structured salvage report includes extracted learnings, guardrails, missing context, local practices, reusable fragments, and restart context to prevent insight loss when drift happens.