qmd-sync

Refresh QMD indexes and embeddings for updated wiki and memory content.

Updated Jun 23, 2026

Search Tags:#embeddings #cron #qmd #reindex #search freshness #agent silos

Ready to run

npx skills add https://github.com/Walliiee/agent-harness --skill qmd-sync

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: qmd-sync
Source: https://github.com/Walliiee/agent-harness/tree/main/skills/qmd-sync
Command: npx skills add https://github.com/Walliiee/agent-harness --skill qmd-sync

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill closes the freshness gap between wiki or memory writes and the vector index, so newly promoted content becomes searchable without waiting for a manual refresh.

Core Features & Use Cases

Cross-agent reindexing: Updates every configured agent silo listed in the shared agents map.
Embedding refresh: Generates missing embeddings after markdown changes, adds, or deletions.
Operational resilience: Supports cron-based runs, scoped manual invocations, skipped silos, and partial-success exits.
Use Case: Run it after a memory promotion so the updated QMD content is discoverable within minutes instead of hours.

Quick Start

Ask the agent to run qmd-sync for the relevant agent silo so recent QMD changes are reindexed and re-embedded.

Frequently Asked Questions about qmd-sync

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

FAQPage Schema

How do I refresh QMD search indexes after updating wiki content?▼

To refresh QMD search indexes, you can trigger a reindexing process that rescans markdown collections and generates missing vectors, allowing newly promoted wiki content to become searchable quickly without waiting for manual updates.

Why does newly added memory content not show up in vector search immediately?▼

Newly added memory content does not appear in vector search immediately due to a freshness gap between content writes and the vector index, which requires missing embeddings to be generated before the updates become searchable.

Can I schedule automatic embedding generation for agent silos?▼

Yes, you can schedule automatic embedding generation for agent silos using cron-driven maintenance runs, which periodically rescan markdown collections and update vectors to maintain search freshness across configured agents.

What is the best way to sync QMD embeddings across multiple specialist agents?▼

The best way to sync QMD embeddings across multiple specialist agents is through cross-agent reindexing, which updates every configured agent silo listed in the shared agents map to ensure consistent search freshness.

How do I handle partial failures when reindexing markdown collections?▼

When reindexing markdown collections, partial failures are handled through operational resilience features that support skipped silos and partial-success exits, ensuring safe skip handling without disrupting the entire synchronization process.

Does qmd-sync work for both manual urgent syncs and scheduled cron maintenance?▼

Yes, qmd-sync works for both manual urgent syncs and scheduled cron maintenance, supporting scoped manual invocations for specific agent silos alongside automated cron-driven runs to refresh QMD indexes and embeddings.