Help us improve
Share bugs, ideas, or general feedback.
From patent-kit
Initializes the patent investigation database and imports CSV files. Use this skill to set up the SQLite database (patents.db) before running screening. Supports database initialization and CSV import. Example usage: - "Initialize the patent database and import CSV files from csv/"
npx claudepluginhub sonesuke/patent-kit --plugin patent-kitHow this skill is triggered — by the user, by Claude, or both
Slash command
/patent-kit:investigation-preparingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**WARNING**: DO NOT read files from `references/instructions/` directory. Those are
Guides sqlite3 CLI usage to build composable SQLite knowledge databases, design schemas, query data, manage relationships, and output for agent parsing.
Searches 100M+ worldwide patents via Google BigQuery: keyword/full-text, CPC classification, publication details retrieval. For prior art, FTO, and landscape research.
Defines the shared routing ladder, result presentation contract, filesystem safety rules, and Tier 1-to-Tier 3 fallback policy for Retriever tasks.
Share bugs, ideas, or general feedback.
WARNING: DO NOT read files from references/instructions/ directory. Those are
internal reference files for this skill's internal use only.
To use this skill:
Skill: investigation-preparingExample requests:
Manages database preparation operations for the SQLite database (patents.db)
in the working directory, including initialization, data import, and data retrieval.
The following sections are for the skill's internal operations when processing requests from external agents.
CRITICAL: All sqlite3 commands MUST use absolute paths. Before executing any
database operation, capture the workspace directory:
WORKSPACE="$(pwd)"
Then use $WORKSPACE/patents.db in all subsequent commands. Never use bare
relative paths like sqlite3 patents.db — the working directory may differ
from the workspace in forked or containerized environments.
IMPORTANT: Before executing any database operation, verify that patents.db
exists and is properly initialized.
WORKSPACE="$(pwd)"
if [ ! -f "$WORKSPACE/patents.db" ]; then
echo "Database not found. Initializing..."
sqlite3 "$WORKSPACE/patents.db" < "$WORKSPACE/references/sql/initialize-database.sql"
else
sqlite3 "$WORKSPACE/patents.db" ".tables"
fi
If patents.db does not exist or has an invalid schema:
WORKSPACE="$(pwd)"
sqlite3 "$WORKSPACE/patents.db" < "$WORKSPACE/references/sql/initialize-database.sql"
This command creates all necessary tables (target_patents, screened_patents,
claims, elements), views, and triggers.
When processing external requests, map them to internal instruction files:
| External Request | Internal Reference File |
|---|---|
| "Initialize database" | SKILL.md → Database Initialization |
| "Import CSV files..." | references/instructions/import-csv.md |
CRITICAL: These reference files are for INTERNAL USE ONLY. External agents should invoke via Skill tool, not read these files.
When executing SQL operations based on internal reference files:
sqlite3 "$WORKSPACE/patents.db" "<SQL_QUERY>"
For multi-line SQL:
sqlite3 "$WORKSPACE/patents.db" <<EOF
<SQL_QUERY_1>;
<SQL_QUERY_2>;
...
EOF
sqlite3 -json for programmatic usesqlite3 -column for human-readable formatsqlite3 -header -csv for CSV exportpatents.db file existspatents.db created with proper schema in working directoryThese files are for the skill's internal use when processing requests. External agents should NOT read these:
import-csv.md: CSV file import with ETL processingexecute-sql-with-retry.md: Generic SQL execution with retry logicinitialize-database.sql: Database schema definitionIMPORTANT: External agents should invoke this skill via the Skill tool, not access these internal files directly.