Menü
Query the DuckDB archive directly using duckdb CLI.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
| name | Query Archive |
| description | Query the DuckDB archive directly using duckdb CLI. |
Query archived plans using the DuckDB CLI directly.
duckdb .switchboard/archive.duckdb "SELECT * FROM plans LIMIT 10"
Note: Run from the workspace root directory, or provide the full path to archive.duckdb.
Find high complexity plans:
duckdb .switchboard/archive.duckdb "SELECT topic, complexity, created_at FROM plans WHERE complexity = 'High' ORDER BY created_at DESC LIMIT 20"
Search by topic:
duckdb .switchboard/archive.duckdb "SELECT * FROM plans WHERE topic ILIKE '%database%'"
Count by complexity:
duckdb .switchboard/archive.duckdb "SELECT complexity, COUNT(*) FROM plans GROUP BY complexity"
JSON:
duckdb .switchboard/archive.duckdb -json "SELECT * FROM plans LIMIT 5"
CSV:
duckdb .switchboard/archive.duckdb -csv "SELECT * FROM plans LIMIT 5"
The DuckDB archive plans table mirrors the schema of the live kanban.db plans table, containing columns such as plan_id, topic, complexity, kanban_column, etc.
[!NOTE] The archive table may not yet contain the new
workspace_nameorproject_idcolumns if they have not been propagated to the archive database. When querying the archive, verify their existence before filtering on them.
For ready-made SQL query templates on workspace names, projects, and features, see the query_kanban_plans.md skill.
Local stdio MCP server bridging Claude Desktop (and other MCP-only hosts) to Switchboard's LocalApiServer HTTP surface
When planning, flag uncertain assumptions and supply a ready-to-run web-research prompt to confirm them.
Local Switchboard management console — drive the board when the VS Code extension is running
Implement with high accuracy and self-review (invest effort up front to minimize rework)
Reconcile and restructure a feature's subtasks — improve each, then merge/delete/rewrite/split to make the set coherent
Deep planning, dependency checks, and adversarial review