swarmvault

Query and maintain SwarmVault knowledge vaults using graph traversal and grounded citations.

Updated Jun 15, 2026
One-click install
npx skills add https://github.com/redlanternstudios/swarmclaw --skill swarmvault-redlanternstudios
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: swarmvault
Source: https://github.com/redlanternstudios/swarmclaw/tree/main/skills/swarmvault
Command: npx skills add https://github.com/redlanternstudios/swarmclaw --skill swarmvault-redlanternstudios

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @swarmvaultai/cli.

What problem does it solve? Working with a SwarmVault knowledge vault requires respecting its three-layer structure (raw/, wiki/, state/) and schema conventions; this Skill teaches the agent to query, ingest, and compile vault content correctly without corrupting compiler-owned files. ## Core Features & Use Cases - Schema-first vault operations: Reads swarmvault.schema.md before any compile or query work and preserves wiki frontmatter fields like page_id, source_ids, and freshness. - Graph-based retrieval: Prefers typed graph queries (query_graph, shortest_path, blast_radius, god_nodes) over grep for relational and dependency questions. - Ingest and maintenance workflow: Adds new sources via ingest_input, re-derives wiki pages with compile_vault, and checks vault health with lint_vault. - Use Case: Ask how two modules in your vault relate, get a grounded answer with source citations, and save it to wiki/outputs/ so it becomes first-class vault content. ## Quick Start Ask the agent to answer a question against the SwarmVault vault and save the grounded answer to wiki/outputs/.

Frequently Asked Questions about swarmvault

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

FAQPage Schema
How do I query a SwarmVault knowledge vault?

Use the query_vault tool for natural-language questions, which returns grounded answers with source_ids and node_ids citations. For relational questions, prefer query_graph, get_node, or shortest_path over text search.

How do I add a new source to a SwarmVault vault?

Call ingest_input with the file path or URL to register it under raw/, then run compile_vault to derive new wiki pages, graph, and search index. Finish with lint_vault to check frontmatter and links.

Can I edit files in the raw/ or wiki/ directories directly?

No. The raw/ directory is immutable and new sources must go through ingest. Wiki pages are compiler-owned, so edits must preserve frontmatter fields like page_id, source_ids, and freshness or the next compile will overwrite them.

What if the SwarmVault MCP server is unavailable?

The same operations are available through the swarmvault CLI or npx -y @swarmvaultai/cli with the working directory set to the vault root, provided the agent has a shell or execute tool.

When should I use graph queries instead of search in a vault?

Use graph queries for relational questions like how X relates to Y or what depends on Z, since the typed graph is more reliable than text search. Read wiki/graph/report.md or wiki/index.md first to understand vault structure.