Query output
Skill SourceWatcher/source-watcher-skills/skills/query-output
Inspect and query SQLite output files produced by Source Watcher pipeline runs. Lists available databases, shows schemas, and runs SQL queries.From its SKILL.md
npx -y skills add SourceWatcher/source-watcher-skills --skill query-outputAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.5 KB, as published. Nobody here has run it
Inspect SQLite output from a Source Watcher pipeline run.
Input: $@
Step 1 - Find available output databases
ls source-watcher-quickstart/source-watcher-api/.source-watcher/*.db 2>/dev/null
If a specific <pipeline-name> or <db-path> was provided, resolve it:
- If it ends in
.db, use it directly as the path - Otherwise, look for
<pipeline-name>.dbunder.source-watcher/
If no .db files exist, tell the user no pipeline output has been produced yet and suggest running a pipeline first via /source-watcher:run-pipeline.
Step 2 - Show available tables
sqlite3 <db-path> ".tables"
List the tables found.
Step 3 - Show schema
For each table, show its schema:
sqlite3 <db-path> ".schema <table-name>"
Step 4 - Run a query
If --query <sql> was provided, run it directly:
sqlite3 -column -header <db-path> "<sql>"
If no query was provided, show a sample of the first table:
sqlite3 -column -header <db-path> "SELECT * FROM <first-table> LIMIT 10;"
Step 5 - Present results
Show the query output in a readable format.
If the result has more than 50 rows, note the truncation and suggest adding a LIMIT clause.
If the user wants to explore further, accept follow-up SQL queries and repeat Step 4.