agent-restore-context

Manage .agent-restore-context-<skill-name> files and verify SessionStart hooks in .claude/settings.json.

154|28|Updated Aug 31, 2018
One-click install
npx skills add https://github.com/commercetools/ui-kit --skill agent-restore-context
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-restore-context
Source: https://github.com/commercetools/ui-kit/tree/main/.agents/skills/agent-restore-context
Command: npx skills add https://github.com/commercetools/ui-kit --skill agent-restore-context

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

When an AI session is cleared or compacted, important in-progress skill state and prompts are lost, interrupting long-running or multi-step agent workflows. This Skill provides a lightweight, file-based mechanism to persist a complete resume prompt so a skill can be reloaded and continue after a context reset.

Core Features & Use Cases

  • Write restore context: Create or overwrite .agent-restore-context-<skill-name> files containing a self-contained resume prompt that tells the agent how to restore state and reload the full skill.
  • Delete cleanup: Remove restore files when workflows are complete; operations are idempotent and safe to repeat.
  • Check & setup validation: Inspect .claude/settings.json and optional settings.local.json for the restore-context SessionStart hook and offer to run setup if missing.
  • Status & previews: List all active restore-context files and show filename-derived skill names with a short content preview to help manage active resumes.
  • Use case: Save progress for long-running tasks like issue implementation or test runs so the agent can resume after a manual /clear or automatic compaction.

Quick Start

Run /agent-restore-context write jira-implement-task -- === RESUMED AFTER CONTEXT RESET === followed by the resume prompt content to save the restore context for the jira-implement-task skill.

Frequently Asked Questions about agent-restore-context

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

FAQPage Schema
How do I restore agent context after a session reset clears my in-progress prompts?

Restoring agent context after a session reset requires persisting a resume prompt to a `.agent-restore-context-<skill-name>` file, which a SessionStart hook automatically injects after `/clear` or `/compact` to reload the skill state.

How do I persist AI agent state across session compaction without losing progress?

Persisting AI agent state across compaction involves writing a self-contained resume prompt to a restore context file in the project root, ensuring the agent can resume long-running workflows like issue implementation or test runs seamlessly.

Where do I configure the hook to automatically reload skill prompts after context loss?

Configuring the hook to automatically reload skill prompts is done in `.claude/settings.json` and optional `settings.local.json`, where you verify the SessionStart hook exists to inject saved restore context files after context loss.

Can I safely delete agent restore context files when a workflow is complete?

You can safely delete agent restore context files when a workflow finishes, as the delete operations are idempotent and can be repeated without errors to clean up `.agent-restore-context-<skill-name>` files in the project root.

What is the best way to check active restore context files and preview their saved skill states?

Checking active restore context files and previewing saved skill states is done by listing all `.agent-restore-context-<skill-name>` files in the project root, which displays filename-derived skill names alongside a short content preview for managing active resumes.

Why does my agent lose its skill prompt state after context compaction?

Agents lose skill prompt state after context compaction because session resets clear in-progress memory, requiring a file-based persistence mechanism to save and inject a self-contained resume prompt via a SessionStart hook.