THE META PROMPT 是一个专业化的提示词仓库,提供经过实战验证的结构化元提示词模板,旨在将通用大语言模型转化为具有增强可靠性和精确度的领域专用推理智能体。与传统提示词库不同,本项目专注于元级别的指令框架,支持系统化任务分解、基于角色的推理以及自我改进的智能体架构。
-
Updated
Feb 7, 2026
#
agent-prompt
Here are 7 public repositories matching this topic...
Real-time Reinforcement Learning & Mirror Therapy built with Gemini 3 Pro.
opencv backend frontend-web api-wrapper gemini-api tts-engine mediapipe q-learning-algorithm genai ai-studio-google agent-prompt gemini-3-pro
-
Updated
Dec 17, 2025 - TypeScript
⛩ 式神召喚術 — Claude Code Skill for generating high-quality agent prompts. Summon shikigami with Shinto-inspired multi-agent workflows, formations, ofuda modules, and taboo systems.
multi-agent claude ai-agent prompt-engineering anthropic agentic-workflow claude-code shikigami agent-prompt claude-code-skill
-
Updated
Mar 20, 2026 - Shell
Trae SKILL：将任何多步骤修复/改进计划转化为完整的代理可执行提示，包括质量检查、原子提交和四引号会话切换。针对TS/npm + DeepSeek harness生态系统进行了优化。
-
Updated
Oct 11, 2026
Portable system prompt that makes any AI agent a topic specialist: task and domain cards, a source ledger, evidence labels and a ten-item quality gate. Thirteen modules, machine-checked.
lint python cli portable harness autonomous-agents ai-agents evidence-based quality-gate llm prompt-engineering ai-governance prompt-testing system-prompt agent-workflow agents-md agent-prompt autoclaw
-
Updated
Nov 7, 2026 - Python
Why your AI agent ignores its system prompt — a Claude Skill that diagnoses before it rewrites. Instruction following, prompt optimization, context engineering, prompt audit. Ships its own eval harnesses.
ai-agents claude instruction-following dspy llm prompt-engineering evals prompt-optimization llm-agents system-prompt system-prompts agent-skills claude-code context-engineering gepa agent-prompt claude-skills claude-plugin claude-code-skill prompt-audit
-
Updated
Aug 31, 2026 - Python
Add this topic to your repo
To associate your repository with the agent-prompt topic, visit your repo's landing page and select "manage topics."