generate_snapshot_batch
Map step — returns one batch of files for the assistant to summarize.
How to use it
generate_snapshot_batch is exposed by the Mason MCP server. Add the server to your MCP client (Claude Desktop, Cursor, Windsurf and others), and the generate_snapshot_batch tool becomes available to the model automatically. See the full listing for setup details and every tool this server provides.
Install Mason
npx -p mason-context mason-mcpOther tools in Mason (17)
Git stats — hot files, stale dirs, commit conventions.
Sync the concept map to Confluence as PM-readable wiki pages.
One-shot orientation for unmapped projects: structure + samples + tests + git.
Smart file previews selected by architectural role.
First call for any task or bug. Matching features + files + tests + blast radius + freshness + recorded decisions, in one call.
Call before editing a file. Traces what's affected — co-change history + references + related tests.
First call for any architecture question. Loads the concept map — feature → file lookup — in one LLM-free call.
Feature-level staleness report — what changed since the snapshot, and whether to refresh incrementally or rebuild.
Marks the project as initialized once the playbook is done.
Start here. Returns the Map-Reduce setup playbook. Idempotent.
Configure Confluence credentials — two-step: list spaces, then persist.
Reduce step — returns every partial + instructions to merge into a unified map.
Record knowledge the code can't express — failed approaches, deprecations, conventions. Git-committed, PR-reviewed, drift-checked.
Persists the partial map for one batch.
Persist the final unified map. Clears partials.
Record verification verdicts — failures flag entries for re-mapping until fixed.
Spot-check map correctness — sampled entries + file skeletons for the assistant to judge, least-recently-verified first.