gaia-refresh-ground-truth

Rescan the filesystem and regenerate a structured diff report against the last ground-truth snapshot.

2|Updated Apr 15, 2026

Search Tags:#memory #ground-truth #incremental-scan #diff-report #filesystem-scan #validator-sidecar

Ready to run

npx skills add https://github.com/gaiastudio-ai/gaia-public --skill gaia-refresh-ground-truth

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: gaia-refresh-ground-truth
Source: https://github.com/gaiastudio-ai/gaia-public/tree/main/plugins/gaia/skills/gaia-refresh-ground-truth
Command: npx skills add https://github.com/gaiastudio-ai/gaia-public --skill gaia-refresh-ground-truth

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires checkpoint.sh, lifecycle-event.sh, and includes scripts (resource) components.

What problem does it solve?

The gaia-refresh-ground-truth skill keeps validator-sidecar ground-truth aligned with the actual project structure by scanning the filesystem, inventorying files, and generating a diff-driven report of changes for review and reconciliation.

Core Features & Use Cases

Automated inventory & diff: detects added, updated, and removed entries by comparing current scans to the last-ground-truth snapshot.
Incremental and full scans: supports both modes and handles first-run initialization, cross-agent synchronization, and dual-write to a committed seed.
Decision logging and governance: writes a diff summary to the decision log and maintains an audit trail for governance and rollback.

Quick Start

Run the gaia-refresh-ground-truth skill to rescan your project and refresh the ground-truth.md in the validator-sidecar.

Frequently Asked Questions about gaia-refresh-ground-truth

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

FAQPage Schema

How do I detect filesystem drift between my project and stored ground-truth?▼

Ground-truth synchronization works by comparing a current filesystem scan against the last committed snapshot. It identifies added, updated, and removed files, then dual-writes the refreshed inventory to a seed file and a decision log for auditability.

When do I need to run a filesystem scan for ground-truth initialization?▼

You need a filesystem scan for ground-truth initialization during first-run setup or cross-agent synchronization. It captures the initial project state and writes a ground-truth.md file with the standard header to establish a baseline snapshot.

Can I perform an incremental scan instead of a full filesystem scan?▼

Yes, you can perform an incremental scan instead of a full filesystem scan. The tool supports both modes, allowing you to efficiently detect recent changes and update the memory-ground-truth format without rescanning unchanged files.

How does the validator-sidecar maintain a decision log for governance?▼

The validator-sidecar maintains a decision log by writing a diff summary of added, updated, and removed entries during each refresh. This dual-write to the committed seed and decision log creates an audit trail for governance and rollback.

What is the best way to keep project memory aligned with actual file changes?▼

The best way to keep project memory aligned is to use an automated inventory tool that rescans the filesystem and regenerates a diff report. This enforces the memory-ground-truth format and dual-writes updates to the seed and decision log.