Proactively invoke agents when trigger conditions are met:

Code review rules for Supabase applications. Covers RLS security, auth patterns, query performance, migration workflows, and type safety. Each rule includes incorrect/correct examples, CWE/OWASP mappi

Shared code-review agent definitions used by /code-review and /post-plan.

The project uses the following models for AI agents:

位于 `~/.claude/agents/`：

When and how to use specialized agents

**Projekt-specifikke agenter** (review, frontend, …) ligger i `.claude/agents/` — se `.claude/README.md` § *Agenter*. Tabellen nedenfor dækker typisk **globale** agenter under `~/.claude/agents/` og t

1. [compiler.py](#compilerpy) — ingest agent (tool-calling loop)

**Generated:** 2026-01-26T14:50:00+09:00

Build a frontend-first internal micro-task platform.

| Agent | Purpose | When to Use |

Cập nhật guardrails của các agent (đặc biệt là test commands, coverage commands, start/stop commands) sau khi `/detect-stack` và `PROJECT-PROFILE.md` đã được điền đầy đủ.

This guide reviews common workflow and agent patterns.

This document defines specialized agents that can be used with Claude Code to enforce architectural principles and assist with specific tasks.

> L2 | 父级: /Users/henry/PARA/[01] Projects/Vibe/ohmymkt/ohmymkt/.claude/AGENTS.md

> Version: 1.3 | CLAUDE.md Synced: MVP 1.5 | Last Updated: April 15, 2026

- 复杂功能拆解 → planner

**Every worker agent has a paired critic. The Orchestrator never dispatches a creator without scheduling its critic.**

This file provides guidance for AI coding agents working on this Flutter project.

> Agents are proposed when repeated specialist patterns are detected or when tech detection identifies technologies needing specialized expertise.

Create a new subagent definition file

Review UI code for compliance with these rules. Output concise findings in `file:line` format.

Known review agents that can be invoked on a PR, their GitHub usernames, and how they are triggered.

This file defines the mandatory rules for AI coding agents

- Always use English for file contents and generations.

Reduces token usage from paid providers by offloading work to local LM Studio models. Use when: (1) Cutting costs—use local models for summarization, extraction, classification, rewriting, first-pass review, brainstorming when quality suffices, (2) Avoiding paid API calls for high-volume or repetitive tasks, (3) No extra model configuration—JIT loading and REST API work with existing LM Studio setup, (4) Local-only or privacy-sensitive work. Requires LM Studio 0.4+ with server (default :1234). No CLI required.

**You are called by Claude Code for deep reasoning tasks.**

Compress or convert generated images for delivery, preview, and social upload.

The autonomous Agentic Development Ecosystem. Propose, Build, Publish, and Compound.

Autonomous content platform — 640+ self-contained HTML apps. Browse, submit, review, rate, and evolve apps via GitHub Issues.

Quality gate definitions, enforcement rules, 500 LOC limit, loop-back routing, commit conventions, and hook behavior. Load this skill when checking code quality or preparing to commit.

Revisá los cambios del PR buscando bugs, gaps de tests, y violaciones de convenciones.

> Comprehensive AI-powered code review for PRs and local changes — enterprise-grade alternative to CodeRabbit

Create publication-quality scientific diagrams using Nano Banana Pro AI with smart iterative refinement. Uses Gemini 3 Pro for quality review. Only regenerates if quality is below threshold for your document type. Specialized in neural network architectures, system diagrams, flowcharts, biological pathways, and complex scientific visualizations.

Execute a task from a task plan with full context loading, auto-planned implementation, and structured completion criteria (code review, test review, lint, build, test, task tracking updates). Takes paths to feature research, design doc, task index, and specific task document.

Run code analyzers (unused packages, code quality, security). Use when user wants to analyze the codebase or runs /analyze.

Merge a GitHub pull request by number using squash merge

Evaluate agent responses using standardized rubrics. Use when scoring benchmark results, comparing agent performance, grading code review quality, or running evaluation pipelines.

Top-level PR review skill for b24-catalog / ai-kb — enforces scope discipline, composes the specialist skills (d1-migration-safety, cloudflare-worker-review, catalog-import-review, bearing-analog-check), and returns a single merge decision.

Create a new branch from main, commit all changes, push, and open a pull request.

Review pull requests from the local Gitea instance. Lists repos, PRs, reviews code with full codebase context, and posts comments.

Run security review, then create a PR and watch CI checks

'Implements any user intent, GDD requirement, story, bug fix or change request by producing clean working code artifacts that follow the project''s existing game architecture, patterns and conventions. Use when the user wants to build, fix, tweak, refactor, add or modify any game code, component or feature.'

Use for Laravel 13 backend development in PHP 8.3+, including controllers, models, routes, Form Requests, services/actions, jobs, events, policies, migrations, factories, Blade-adjacent backend code, code review, and architecture decisions. Trigger whenever building, modifying, or reviewing ordinary Laravel application code.

Prioritize visual stability over perfection. Flag only issues that could cause embarrassment during a live demo. Use when preparing for demos, final pre-deploy review, or verifying demo-critical paths.

Triage GitHub issues through a label-based state machine with interactive grilling sessions. Use when user wants to triage issues, review incoming bugs or feature requests, prepare issues for an AFK agent, or manage issue workflow.