Menü
"Parallel SQL-query audit, one agent per .py file."
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 | reviewQueries |
| description | "Parallel SQL-query audit, one agent per .py file." |
Launch parallel queries-reviewer agents to audit SQL query functions for efficiency, reusability, dead code, and quality. One agent per .py file.
$ARGUMENTS
CRITICAL EXECUTION RULE — follow this EXACTLY:
Determine the scope based on $ARGUMENTS:
A) No arguments (empty or blank): use Glob to list all .py files in backend/database/queries/ (exclude __init__.py). Each file becomes a target.
B) One or more .py file paths (space-separated, e.g., backend/database/queries/accounts.py backend/database/queries/auth.py): use those files as targets.
C) The word diffs: run git diff --name-only and git diff --cached --name-only via Bash. From the combined output, keep only files that match backend/database/queries/*.py (exclude __init__.py). Each matching file becomes a target. If no query files appear in the diffs, inform the user "No query files found in current diffs." and STOP.
Launch exactly one queries-reviewer agent per target file, all in parallel in a single message with multiple Agent tool calls. Each agent uses run_in_background: true and subagent_type: "queries-reviewer". Each agent's prompt: "Analyze ONLY the file {file_path}. Focus exclusively on this file and on the usage flow of the query functions it contains. Do NOT read or analyze any other query file in backend/database/queries/."
After launching, your response MUST END IMMEDIATELY. Write a brief message listing the agents launched (count and file names) and tell the user you will report when they finish — then STOP. Do NOT add any more tool calls.
Do NOT poll, retry, resume, or call any tool after launching the agents. The main thread must be completely idle until the automatic completion notifications arrive.
Only when ALL agents have completed (via automatic notifications), present a unified summary to the user with the results from each file.
"Parallel error-handling audit across the given directories."
"Parallel test-quality and coverage audit across the given directories."
"Parallel test runner across the given directories."
Ship a feature from a worktree: commit, push, create PR, merge to master, clean up the worktree and branches, then rebase all remaining worktrees with conflict resolution. Use when user says "ship", "ship it", "merge this feature", "send to master", or wants to finalize a worktree and sync the rest.
Diseña un plan para implementar una funcionalidad y lo devuelve para que el usuario lo apruebe antes de ejecutarlo. Los dos últimos pasos del plan son siempre, en este orden, la invocación de @agent-tests-author-from-diff y @agent-guides-updater-from-diff.
Full pre-commit review covering both backend and frontend in parallel. Runs the backend architectural review (spawning background subagents across layers, tests, docs, and queries) and the frontend CLAUDE.md compliance review at the same time, then merges both into one consolidated report.
