// Use when working with a SwarmVault knowledge vault (raw/, wiki/, swarmvault.schema.md). Establishes schema-first conventions and prefers graph queries over broad search.
AI agent runtime and multi-agent orchestration platform. Teaches agents how to use SwarmClaw's 6 primitive tools, persistent memory, dreaming, delegation, connectors, credentials, and the skill system. Use when an agent is running on SwarmClaw and needs to understand the platform's capabilities.
Manage your SwarmClaw agent fleet — agents, tasks, chats, chatrooms, goals, schedules, memory, wallets, connectors, autonomy, and 40+ more command groups. Use when asked to dispatch work, check agent status, coordinate multi-agent work, run diagnostics, manage schedules, set goals, or orchestrate across a SwarmClaw dashboard instance.
Delegate coding tasks to external coding agents (Claude Code, Codex, Pi, OpenCode) via shell. Use when: (1) building new features or apps in a separate project, (2) reviewing PRs, (3) refactoring large codebases, (4) iterative coding that needs file exploration. NOT for: simple one-liner fixes (just edit directly), reading code (use read/file tools), or work inside the SwarmClaw workspace itself.
GitHub operations via `gh` CLI: issues, PRs, CI runs, code review, API queries. Use when: (1) checking PR status or CI, (2) creating/commenting on issues, (3) listing/filtering PRs or issues, (4) viewing run logs. NOT for: local git operations (use git directly), non-GitHub repos, or cloning (use git clone).
Use Google Workspace CLI (`gws`) for Drive, Docs, Sheets, Gmail, Calendar, Chat, and related Workspace API tasks.
Generate or edit images via Gemini 3 Pro Image (Nano Banana Pro). Use when asked to create, generate, or edit images and a Gemini API key is available. Supports text-to-image generation, single-image editing, and multi-image composition (up to 14 images).
| name | swarmvault |
| description | Use when working with a SwarmVault knowledge vault (raw/, wiki/, swarmvault.schema.md). Establishes schema-first conventions and prefers graph queries over broad search. |
| homepage | https://swarmvault.ai |
| metadata | {"openclaw":{"capabilities":["knowledge-base","knowledge-graph","retrieval","vault"],"requires":{"bins":["npx"]}}} |
Use when the agent has a SwarmVault MCP server enabled (transport stdio, command npx -y @swarmvaultai/cli mcp) pointed at a vault directory.
A SwarmVault workspace is a three-layer knowledge system:
raw/ — immutable source inputs (PDFs, transcripts, code, emails, URLs, sheets). Never edit.wiki/ — generated markdown owned by the agent and the SwarmVault compiler. Pages carry frontmatter (page_id, source_ids, node_ids, freshness, source_hashes).state/ — generated indexes, graphs, and approvals. Treat as opaque output of compile.The vault contract lives in swarmvault.schema.md at the workspace root. The vault config lives in swarmvault.config.json.
swarmvault.schema.md first before any compile or query work. It defines categories, naming, freshness rules, and grounding conventions for this specific vault.wiki/graph/report.md before broad file searching when it exists; otherwise start with wiki/index.md. Both summarize the vault structure so you don't re-scan everything.raw/ as immutable. Never edit, rename, or delete files there. New sources go through ingest.wiki/ as compiler-owned. Edits should preserve frontmatter fields exactly: page_id, source_ids, node_ids, freshness, source_hashes. If those drift, the next compile will overwrite or flag the page.wiki/outputs/ (use the query or explore tools) instead of leaving them only in chat. That way they become first-class vault content for next time.The SwarmVault MCP server exposes the following tools (names are prefixed by SwarmClaw with mcp_<sanitized server name>_, e.g. mcp_SwarmVault_query_vault). Match the user's intent to the closest tool:
Vault inspection:
workspace_info — return current vault paths and high-level counts. Use this first when you've never seen this vault.list_sources — list source manifests under raw/.search_pages — full-text search across compiled wiki pages.read_page — read a specific wiki page by its wiki/-relative path.Graph (prefer over grep for relational questions):
graph_report — machine-readable graph report and trust artifact. Read this before broad searching.query_graph — traverse the graph from search seeds without calling an LLM provider.get_node — explain a graph node, its page, community, neighbors, and group patterns.get_neighbors — neighbors of a node or page target.get_hyperedges — list graph hyperedges, optionally filtered.shortest_path — shortest path between two graph targets.god_nodes — highest-connectivity nodes (the vault's hubs).blast_radius — impact analysis: what depends on this file or module?Question answering:
query_vault — natural-language question against the vault. Returns grounded citations. Pass save: true to persist the answer to wiki/outputs/.Ingest and maintenance:
ingest_input — add a file path or URL to raw/ and register it as a managed source.compile_vault — re-derive wiki/ pages, graph, and search index. Run after ingest, after schema changes, or when freshness is stale.lint_vault — anti-drift and vault health checks.If the MCP server is unavailable but the agent has a shell or execute tool, the same operations are available via swarmvault <subcommand> (or npx -y @swarmvaultai/cli <subcommand>) with the working directory set to the vault root.
For a fresh question against the vault:
workspace_info if you haven't already, then read swarmvault.schema.md. If wiki/graph/report.md or wiki/index.md exists, skim it.query_vault (or query_graph / get_node / shortest_path for relational questions). Cite returned source_ids and node_ids.ingest_input for the missing source, then compile_vault.query_vault save: true so it becomes vault content under wiki/outputs/.For a new source the user mentions:
ingest_input the file/URL.compile_vault to derive new wiki pages, graph, and search index.lint_vault to check frontmatter and links.wiki/sources/ and confirm provenance.compile against an unreviewed change to swarmvault.schema.md — lint first.wiki/candidates/) to wiki/concepts/ or wiki/entities/ without the user's confirmation; the approval flow exists for a reason.