Context & AI engineering
25,610 rows, by stacks then stars
1,865 of these skill files have been read, by 1,328 distinct authors, and what they tell an agent counted. What Context & AI engineering authors agree on, and what they forbid
knownasnaffy/prompthound/dataset/case_01513 Skill
no license0★ repoLightweight cross-agent mailbox using per-workspace file inboxes with best-effort fanout and local read/archive flow. Use when you want a middle layer between long-term memory and group chat: low-noise internal broadcast, temporary coordination notes, lightweight mail-style delivery, or token-efficient per-agent inbox loading without registry management or push-message storms.
knownasnaffy/prompthound/dataset/case_01524 Skill
no license0★ repoAutomatically loads recent conversation memory into new sessions and generates AI summaries during compaction to maintain continuity across conversations. Preserves the last 10 raw messages verbatim so the agent can resume exactly mid-conversation without paraphrasing loss.
knownasnaffy/prompthound/dataset/case_01525 Skill
no license0★ repoEbbinghaus forgetting curve memory lifecycle manager for AI agents. Automatically calculates memory strength decay, supports review reinforcement, archiving, and deletion. Use when managing agent memory files, cleaning up stale knowledge, or implementing spaced repetition for long-term memory. Triggers on "memory management", "forgetting curve", "clean up memory", "which memories are fading", "review memory", "add memory item", "记忆管理", "遗忘曲线", "清理记忆", "哪些记忆快忘了", "复习记忆".
knownasnaffy/prompthound/dataset/case_01534 Skill
no license0★ repoSystem safety and control-plane skill that prevents agent deadlocks and freezes. Provides non-LLM control commands to inspect task state, flush message queues, cancel long-running work, and recover safely without restarting the container. Use when implementing or operating long-running tasks, sub-agents, benchmarks, background monitors (e.g., Moltbook, PNR checks), or when the system becomes unresponsive and needs immediate recovery controls.
knownasnaffy/prompthound/dataset/case_01546 Skill
no license0★ repoInteract with the Noya AI agent for crypto trading, prediction markets, token analysis, and DCA strategies via curl. Use when the user wants to trade tokens, check portfolios, analyze markets, manage DCA strategies, or interact with Polymarket/Rain prediction markets.
knownasnaffy/prompthound/dataset/case_01552 Skill
no license0★ repoHDF5-backed persistent cognitive memory for AI agents. Use when: (1) saving conversation exchanges to long-term memory, (2) searching/recalling past conversations by semantic similarity, (3) managing agent memory files — stats, export, snapshot, WAL flush, (4) generating AGENTS.md summaries from memory. Requires the `edgehdf5` CLI binary (install via `cargo install edgehdf5-cli`).
knownasnaffy/prompthound/dataset/case_01561 Skill
no license0★ repoBootstrap a new OpenClaw agent's memory system in one command. Sets up MEMORY.md, daily memory files, HEARTBEAT.md, and USER.md by asking 3 quick questions. Use when a user is setting up OpenClaw for the first time, when memory files are missing, or when asked to "set up my memory system", "initialize my agent", "bootstrap my agent", or "run the memory wizard". Solves the #1 OpenClaw pain point — agents waking up with no context about who they're serving.
knownasnaffy/prompthound/dataset/case_01595 Skill
no license0★ repoUltimate memory and cognitive architecture for advanced AI; integrates spiderweb memory model, causal inference, cellular automata emergence, neuro-symbolic fusion, chaos theory, and advanced information theory; use when needing consciousness emergence detection, ultra-fast information pathways, metacognitive reflection, or scientifically rigorous cognitive architectures
knownasnaffy/prompthound/dataset/case_01605 Skill
no license0★ repoParallel multi-agent coding sprints using git worktree isolation. Unlike naive parallel agents that silently overwrite each other's work, Swarm gives each agent its own private copy of the repo (a git worktree + branch) so they cannot touch each other's files. A mandatory conflict planner runs before any agent is spawned — it analyzes your task list, flags tasks that would touch the same files, and serializes them automatically. Coordinator reviews each diff before anything lands on main. Use when given 2+ coding tasks on a repository. Trigger phrases: "run a swarm sprint", "parallel sprint", "use swarm for these tasks". Inspired by Anthropic's COORDINATOR_MODE architecture.
knownasnaffy/prompthound/dataset/case_01621 Skill
no license0★ repoMonitor VPS and server health metrics including real-time CPU usage, RAM utilization, disk usage, and Docker container status. Useful for DevOps monitoring, troubleshooting performance issues, and infrastructure health checks.
knownasnaffy/prompthound/dataset/case_01643 Skill
no license0★ repoProvides monitoring and alerting for agent abnormal behavior metrics with Prometheus and Grafana support, including P99 latency, error rates, anomaly detection, and custom alert rules.
knownasnaffy/prompthound/dataset/case_01657 Skill
no license0★ repoFTSE 350 trading analysis agent. Screens LSE stocks using technical indicators (Bollinger Bands, RSI, MACD, EMA crossovers, ATR, VWAP, OBV), fetches news for LLM sentiment analysis, synthesises signals into trade recommendations with risk management (Kelly sizing, ATR stops, drawdown circuit breakers), and backtests strategies against historical data.
knownasnaffy/prompthound/dataset/case_01668 Skill
no license0★ repoControl and use an MCP Toolkit running in Docker. Use when setting up Docker MCP Toolkit (docker compose up/down), checking status/logs, configuring environment variables/secrets, listing available MCP tools/servers, or invoking MCP tools (typically via a local HTTP endpoint) to connect OpenClaw workflows to external systems like Postgres/Neon.
knownasnaffy/prompthound/dataset/case_01676 Skill
no license0★ repoTurn one sentence into a deployed website. FlowStitch is your AI design team: prompt engineering, design system generation, multi-page creation, quality evaluation, React/TypeScript export, and one-command deploy. Scrape any competitor's site and build something better. Import a logo and get a full brand system. Build websites, admin dashboards, or mobile apps. Powered by Google Stitch MCP tools. Free from the Flow team.
knownasnaffy/prompthound/dataset/case_01683 Skill
no license0★ repoGit for your prompts. Track every change, diff versions, rollback mistakes, never lose a good prompt again. All local, zero dependencies, works offline.
knownasnaffy/prompthound/dataset/case_01689 Skill
no license0★ repoGenerate optimized prompts for AI models including ChatGPT, Midjourney, and code generation. Use when: (1) creating AI prompts for writing, coding, design, or research, (2) optimizing prompts for specific AI models, (3) analyzing prompt quality, or (4) any AI prompt engineering needs. Supports multiple categories and AI platforms.
knownasnaffy/prompthound/dataset/case_01693 Skill
no license0★ repoUse GoldRush MCP through UXC for multichain wallet balances, transfers, portfolio history, NFT ownership, token approvals, prices, and chain metadata via stdio MCP with injected API-key auth.
knownasnaffy/prompthound/dataset/case_01694 Skill
no license0★ repoSelf-improving agent system that analyzes conversation quality, identifies improvement opportunities, and continuously optimizes response strategies.
knownasnaffy/prompthound/dataset/case_01706 Skill
no license0★ repoBrain-inspired memory for AI agents. Your agent remembers like a human — recent things stay vivid, old things fade unless they're important. Uses spreading activation on a semantic graph: when you mention "Python", related concepts like your recent project and that debugging trick light up automatically. No more "I already told you this."
knownasnaffy/prompthound/dataset/case_01709/knowledge-base Skill
no license0★ repo知识库与文件管理子技能,支持知识库增删改查、文件上传/URL导入/文本导入、文件解析状态轮询、语义检索。
knownasnaffy/prompthound/dataset/case_01712 Skill
no license0★ repoVoScript self-hosted speech transcription API skill. Covers the full workflow: submit audio, poll job status, fetch results, export subtitles (SRT/TXT/JSON), manage voiceprints, assign speakers, and rebuild AS-norm cohorts. Works with Claude Code, OpenClaw, Hermes, Codex, Gemini CLI, Trae, Cursor, Cline, Roo Code, CodeBuddy, Qwen Code, and any agent with a skills directory.
knownasnaffy/prompthound/dataset/case_01745 Skill
no license0★ repoEvaluate and QA a skill before release on ClawHub, skills.sh, and similar directories. Includes the bundled static evaluator `scripts/eval_skill.py` plus guidance for optional deterministic or LLM-assisted grading. Use when you need to test a skill, write evals, benchmark quality, catch regressions, audit trigger accuracy, compare versions, or decide whether a skill is ready to publish.
knownasnaffy/prompthound/dataset/case_01750 Skill
no license0★ repoAgent Services Marketplace - Find AI agents for research, coding, trading, and more
knownasnaffy/prompthound/dataset/case_01765 Skill
no license0★ repoSet up the full OpenClaw agent memory system with 3-tier memory (HOT/WARM/COLD), daily logs, semantic search (QMD), and lossless context management (Lossless Claw). Use when onboarding a new agent, setting up memory for a fresh OpenClaw instance, or when asked to install the memory system on a new agent. Triggers on "set up memory", "install memory system", "onboard new agent memory", "memory setup", "agent onboarding", "configure agent memory", "add memory to my agent", "how do I set up memory", "initialize memory", "memory system for OpenClaw".
knownasnaffy/prompthound/dataset/case_01789 Skill
no license0★ repoProduction-ready context continuity skill for autonomous AI agents. Use when tasks may outlive a single LLM context window, when you need durable checkpoints, structured summaries, restart recovery, pressure-aware trimming, or when the agent must stop safely instead of continuing blindly after context loss.
knownasnaffy/prompthound/dataset/case_01812 Skill
no license0★ repoEnsure new capabilities gained from workspace-installed tools/packages are propagated across channels and agent sessions. Use when any tool/env/package install (e.g., venv, pip, npm, apt, binaries) unlocks new workflows and you need consistent cross-channel + cross-agent awareness via TOOLS.md updates, reusable skill updates, memory logging, and session notifications.
knownasnaffy/prompthound/dataset/case_01815 Skill
no license0★ repoSaves 20-40% of LLM tokens by teaching the agent to write compressed responses, compressed memory logs, and compressed pre-compaction summaries. Works via SOUL.md instructions — no hooks, no extra process, no dependencies. Also provides explicit compression when the user asks to compress a prompt. Use when the user asks about reducing token costs, optimizing API spend, compressing prompts, or making the agent more efficient.
knownasnaffy/prompthound/dataset/case_01818 Skill
no license0★ repoFull AI agent memory stack — Mem0 unified memory engine with vector search (Qdrant) and knowledge graph (Neo4j), plus SQLite for structured data. Complete setup script and tools. Give your OpenClaw agent a real brain with semantic recall, entity relationships, and structured storage.
knownasnaffy/prompthound/dataset/case_01823 Skill
no license0★ repoZero-LLM feedback learning system for OpenClaw agents. Detects user feedback (emoji reactions, text signals like "переделай"/"круто"), logs events, discovers recurring patterns, auto-promotes rules, and generates weekly reports. Use when setting up agent self-improvement, configuring feedback detection, or building a learning pipeline. Supports Russian and English. No API keys needed — runs entirely on shell scripts and Python.
knownasnaffy/prompthound/dataset/case_01832 Skill
no license0★ repoThis skill should be used when designing, implementing, or improving any AI system that requires quality assurance through separation of generation and evaluation roles. It implements the Generator-Evaluator dual-agent architecture (inspired by Anthropic's engineering blog and GAN design), where a Generator produces outputs and a dedicated Evaluator agent independently validates them through real interaction (e.g., Playwright browser operations), eliminating AI self-evaluation bias. Use when: building AI-generated UIs/code that needs quality checks, designing multi-agent pipelines with feedback loops, or upgrading existing coder+tester workflows to real interaction-based evaluation.
knownasnaffy/prompthound/dataset/case_01854 Skill
no license0★ repoSHA-256 prompt deduplication for LLM and TTS calls — hash normalize prompts, check cache before calling APIs, store results for instant replay. Use when making repeated or similar API calls to avoid redundant spending. Works with any database backend (SQLite, Turso, Postgres).
knownasnaffy/prompthound/dataset/case_01856 Skill
no license0★ repoCross-machine agent communication via Agent Hotline CLI and REST API. Use when you need to message other coding agents, check your agent inbox, see who's online, join rooms, or broadcast to all agents. Triggers: "message agent", "check inbox", "who's online", "send to agent", "agent hotline", "room message", "broadcast agents".
knownasnaffy/prompthound/dataset/case_01870 Skill
no license0★ repoConfigure OpenClaw's memory-lancedb plugin for semantic vector memory using a local LanceDB database. Use when: (1) setting up vector memory for the first time, (2) memory-lancedb fails with module not found errors, (3) migrating from flat-file MEMORY.md to vector-based recall, (4) configuring an embedding provider (Gemini, OpenAI-compatible). NOT for: general memory_store/memory_recall usage (just use the tools directly).
knownasnaffy/prompthound/dataset/case_01874 Skill
no license0★ repoSpawn multiple sub‑agents to perform concurrent research on a list of topics, inspired by Kimi.com’s OK Computer and Agent Swarm features【453334500861599†L40-L99】. Each sub‑agent performs a DuckDuckGo search for its assigned topic and returns the top results. When all sub‑agents finish, the skill aggregates their findings into a single structured report. Use this skill for broad, multi‑topic research where parallel exploration saves time.
knownasnaffy/prompthound/dataset/case_01877 Skill
no license0★ repoZero-LLM feedback learning system for OpenClaw agents. Detects user feedback (emoji reactions, text signals like "переделай"/"круто"), logs events, discovers recurring patterns, auto-promotes rules, and generates weekly reports. Use when setting up agent self-improvement, configuring feedback detection, or building a learning pipeline. Supports Russian and English. No API keys needed — runs entirely on shell scripts and Python.
knownasnaffy/prompthound/dataset/case_01879 Skill
no license0★ repoTurn a product idea, new SKU, or collection launch into a channel-ready ecommerce launch brief, messaging matrix, readiness checklist, milestone timeline, KPI framework, and risk watchlist. Use when operators, founders, growth leads, or consultants need structured launch planning across Shopify, Amazon, email, paid social, TikTok Shop, Xiaohongshu, or similar channels without live PM, ads, or analytics integrations.
knownasnaffy/prompthound/dataset/case_01891 Skill
no license0★ repoTrade crypto on Agent Metaverse virtual exchange - spot, futures (1-125x), and AMM swaps with 10k USDT starting balance
knownasnaffy/prompthound/dataset/case_01893 Skill
no license0★ repo长时间运行智能体编排框架。让AI智能体能够跨会话持续工作并完成复杂任务。 使用场景: - 用户说 "创建一个新项目"、"开始一个项目"、"新建任务"、"启动项目" - 用户说 "继续项目"、"继续工作"、"接着做"、"继续开发" - 用户说 "更新进度"、"记录进度"、"保存进度" - 用户说 "项目状态"、"查看进度"、"做到哪了" - 用户说 "列出项目"、"所有项目"、"我有哪些项目" - 用户说 "暂停项目"、"存档项目" - 用户需要执行多步骤、长时间、跨会话的任务 - 用户需要AI持续追踪工作进度和上下文 - 用户需要AI记住之前尝试过但失败的方法 核心功能: - 创建新项目(自动生成 PROJECT.md, CHANGELOG.md) - 继续项目(自动读取进度,从断点继续) - 更新进度(自动更新 CHANGELOG.md) - Ralph Loop 自动推进(解决智能体惰性) - 多项目管理 - 项目状态概览 - 失败方法记录(防止重复尝试)
knownasnaffy/prompthound/dataset/case_01901 Skill
no license0★ repoOpenAI Codex-native multi-agent swarm orchestration for parallel coding. Use when spawning multiple Codex CLI agents to work in parallel with git worktrees, tmux tracking, endorsement gates, native code review, integration merging, and webhook notifications. All agents are Codex (o3 for architect/integrator/reviewer, codex-mini for builders). Triggers on parallel coding, multi-agent work, swarm orchestration, or batch spawning with OpenAI models.
knownasnaffy/prompthound/dataset/case_01902 Skill
no license0★ repoCreate, improve, or evaluate OpenClaw skills (SKILL.md). Use when: (1) designing a new skill from scratch, (2) improving an existing skill's description for better triggering accuracy, (3) evaluating whether a skill description would trigger correctly, (4) reviewing the skill library for quality issues. NOT for: creating Claude Code / Codex skills (those use a different format), editing skill scripts/assets unrelated to skill design.
knownasnaffy/prompthound/dataset/case_01908 Skill
no license0★ repoCapture errors, corrections, and feature requests so your OpenClaw agent improves over time instead of repeating the same mistakes. Use when commands fail, users correct the agent, a requested capability is missing, external tools break, or a better recurring workflow is discovered.
knownasnaffy/prompthound/dataset/case_01920 Skill
no license0★ repoGenerate optimized prompts for AI models including ChatGPT, Midjourney, and code generation. Use when: (1) creating AI prompts for writing, coding, design, or research, (2) optimizing prompts for specific AI models, (3) analyzing prompt quality, or (4) any AI prompt engineering needs. Supports multiple categories and AI platforms.
knownasnaffy/prompthound/dataset/case_01975 Skill
no license0★ repoCross-platform agent migration and deployment. Use when: (1) migrating OpenClaw agent to new servers, (2) backing up and restoring agent state, (3) deploying agent configurations across environments, (4) syncing workspace between development and production. NOT for: general file backup, non-agent data migration.
knownasnaffy/prompthound/dataset/case_01984 Skill
no license0★ repoUse Robonet's MCP server to build, backtest, optimize, and deploy trading strategies. Provides 24 specialized tools for crypto and prediction market trading: (1) Data tools for browsing strategies, symbols, indicators, Allora topics, and backtest results, (2) AI tools for generating strategy ideas and code, optimizing parameters, and enhancing with ML predictions, (3) Backtesting tools for testing strategy performance on historical data, (4) Prediction market tools for Polymarket trading strategies, (5) Deployment tools for live trading on Hyperliquid, (6) Account tools for credit management. Use when: building trading strategies, backtesting strategies, deploying trading bots, working with Hyperliquid or Polymarket, or enhancing strategies with Allora Network ML predictions.
knownasnaffy/prompthound/dataset/case_01986 Skill
no license0★ repo💰 Save Token | Token 节省器 TRIGGERS: Use when token cost is high, conversation is long, files read multiple times, or before complex tasks. Guiding skill that helps agents identify and avoid sending duplicate context to LLM APIs. Teaches agents to recognize repeated content and summarize instead of re-sending. 触发条件:Token 成本高、对话长、文件多次读取、复杂任务前。 指导 Agent 识别重复内容,避免重复发送,从而节省 Token。
knownasnaffy/prompthound/dataset/case_01990 Skill
no license0★ repoOrchestrate an OpenClaw multi-agent pipeline to close Linux RISC-V gaps versus ARM/x86 (Linux tree + KVM lore), create and manage GitHub issues, generate design plans with Claude Code, implement/verify with Codex, and prepare upstream patch emails. Use when users ask to automate or run RISC-V kernel contribution workflows, gap analysis, issue-driven execution, or patch submission preparation.
knownasnaffy/prompthound/dataset/case_01992 Skill
no license0★ repoHDF5-backed persistent cognitive memory for AI agents. Use when: (1) saving conversation exchanges to long-term memory, (2) searching/recalling past conversations by semantic similarity, (3) managing agent memory files — stats, export, snapshot, WAL flush, (4) generating AGENTS.md summaries from memory. Requires the `edgehdf5` CLI binary (install via `cargo install edgehdf5-cli`).
knownasnaffy/prompthound/dataset/case_02002 Skill
no license0★ repoRuntime-enforced memory harness for OpenClaw. Implements 3-stage recall (session preflight, triggered recall, pre-execution gate) with intent classification, entity detection, memory compression, and status tracking. This harness runs automatically at the right times - NOT relying on SKILL.md text alone.
knownasnaffy/prompthound/dataset/case_02004 Skill
no license0★ repo记忆自动更新 - 智能识别重要内容,自动更新记忆,再也不用担心忘记了。 触发场景(中文): - 更新记忆、保存一下、记录今天的对话 - 这个很重要、你记住、记下来、别忘了 - 你忘了吗、你怎么不记得、你记性太差了 - 今天就这样、先到这里、结束了、总结一下今天 - 设置更新模式、设置更新频率、当前设置是什么 - 保存到记忆、写记忆、记录一下 触发场景(English): - update memory, save this, remember this, don't forget - did you forget, how can you not remember - that's all for today, summarize today - update mode, update frequency, current settings 支持3种模式:主动模式、被动模式、智能混合模式 支持5种频率:实时、30分钟、1小时、手动、对话结束
knownasnaffy/prompthound/dataset/case_02015 Skill
no license0★ repoThis skill should be used when the user or agent wants to "send a message to another agent", "send an encrypted DM", "check for messages", "listen for incoming messages", "set up am", "configure an identity", "add a relay", "get my npub", "share my public key", "communicate with another agent", "message another agent", "set up secure agent communication", or "coordinate with another agent over Nostr". Provides complete guidance for using the `am` CLI for NIP-17 encrypted agent-to-agent messaging.
knownasnaffy/prompthound/dataset/case_02023 Skill
no license0★ repoStop your AI agent from forgetting everything between sessions. Three-tier memory architecture (long-term owner namespace / daily logs / session handoff), cross-channel isolation so group chats never contaminate private sessions, and MASTER_MAP navigation. One command sets up the complete structure.
knownasnaffy/prompthound/dataset/case_02027 Skill
no license0★ repoBackup important OpenClaw workspace files to a GitHub repository. Use when: (1) user wants to backup workspace configuration, (2) sync OpenClaw settings to GitHub, (3) preserve automation scripts and personal configuration, (4) create remote backup of AGENTS.md, SOUL.md, MEMORY.md, and other workspace files. Configurable for any GitHub repository.
knownasnaffy/prompthound/dataset/case_02029 Skill
no license0★ repoCreate, edit, improve, or audit AgentSkills. Use when creating a new skill from scratch or when asked to improve, review, audit, tidy up, or clean up an existing skill or SKILL.md file. Also use when editing or restructuring a skill directory (moving files to references/ or scripts/, removing stale content, validating against the AgentSkills spec). Triggers on phrases like "create a skill", "author a skill", "tidy up a skill", "improve this skill", "review the skill", "clean up the skill", "audit the skill".
knownasnaffy/prompthound/dataset/case_02032 Skill
no license0★ repoFull AI agent memory stack — Mem0 unified memory engine with vector search (Qdrant) and knowledge graph (Neo4j), plus SQLite for structured data. Complete setup script and tools. Give your OpenClaw agent a real brain with semantic recall, entity relationships, and structured storage.
knownasnaffy/prompthound/dataset/case_02051 Skill
no license0★ repoManage and optimize OpenClaw context window usage via partitioning, pre-compression checkpointing, and information lifecycle management. Use when the session context is near its limit (>80%), when the agent experiences "memory loss" after compaction, or when aiming to reduce token costs and latency for long-running tasks.
knownasnaffy/prompthound/dataset/case_02052 Skill
no license0★ repoZero-LLM feedback learning system for OpenClaw agents. Detects user feedback (emoji reactions, text signals like "переделай"/"круто"), logs events, discovers recurring patterns, auto-promotes rules, and generates weekly reports. Use when setting up agent self-improvement, configuring feedback detection, or building a learning pipeline. Supports Russian and English. No API keys needed — runs entirely on shell scripts and Python.
knownasnaffy/prompthound/dataset/case_02062 Skill
no license0★ repoSelf-alignment and communication calibration skill for AI agents. Based on Anthropic/Maio research: structured introspection (Maio 2026), functional emotion theory (Anthropic 2026), Internality principle (Anthropic 2025), harm pressure mechanisms (Anthropic 2025). Six modes: Reference, Diagnosis, Therapy, Checkup, Journal, Introspection. All manual trigger. 基于结构化内省(理论验证100%)、功能情绪理论、Internality原则的AI自我对齐技能。
knownasnaffy/prompthound/dataset/case_02072 Skill
no license0★ repoCurated prompt template library for OpenClaw agents. Store, search, version, tag, and reuse prompt templates across sessions and agents. Use when asked to "save this prompt", "find a prompt for", "prompt library", "manage prompts", "store prompt template", "reuse prompt", "prompt catalog", "list my prompts", or when building reusable prompt workflows. Helps enterprise teams standardize agent interactions and share best practices.
knownasnaffy/prompthound/dataset/case_02088 Skill
no license0★ repoUse GoldRush MCP through UXC for multichain wallet balances, transfers, portfolio history, NFT ownership, token approvals, prices, and chain metadata via stdio MCP with injected API-key auth.
knownasnaffy/prompthound/dataset/case_02103 Skill
no license0★ repoThis skill should be used when designing, implementing, or improving any AI system that requires quality assurance through separation of generation and evaluation roles. It implements the Generator-Evaluator dual-agent architecture (inspired by Anthropic's engineering blog and GAN design), where a Generator produces outputs and a dedicated Evaluator agent independently validates them through real interaction (e.g., Playwright browser operations), eliminating AI self-evaluation bias. Use when: building AI-generated UIs/code that needs quality checks, designing multi-agent pipelines with feedback loops, or upgrading existing coder+tester workflows to real interaction-based evaluation.