Name: ai_agents_frameworks
Author: Wang200935

Imported from Wang200935/openclaw-skills (misc/_merge_backups/auto_20260315_070002/misc/ai/ai-agents-frameworks/SKILL.md). Install upstream with npx skills add Wang200935/openclaw-skills --skill ai-agents-frameworks. Copyright stays with the author.

Agent Framework Patterns

Use this skill for agentic systems.

Core rules

Keep control loops explicit.
Make tool boundaries auditable.
Separate planner state from executor state.
Add guardrails around destructive actions.
Prefer deterministic repair flows over free-form autonomy for ops tasks.

Common pattern

observe -> diagnose -> decide -> act -> verify

For ops agents

Log every action.
Require confirmation for dangerous steps unless already authorized.
Make retries bounded and visible.

MERGED_FROM_BACKUP: C:\Users\wang.openclaw\skills\ai\ai-agents-frameworks\SKILL.md

name: ai_agents_frameworks description: Reason about multi-step agent systems, tool use, orchestration, and control loops using concepts from LangGraph, CrewAI, tool-calling examples, and agent courses. Use when building AI agents that plan, call tools, manage state, or coordinate subtasks.

Agent Framework Patterns

Use this skill for agentic systems.

Core rules

Keep control loops explicit.
Make tool boundaries auditable.
Separate planner state from executor state.
Add guardrails around destructive actions.
Prefer deterministic repair flows over free-form autonomy for ops tasks.

Common pattern

observe -> diagnose -> decide -> act -> verify

For ops agents

Log every action.
Require confirmation for dangerous steps unless already authorized.
Make retries bounded and visible.

{ "ocm": "1", "id": "wang200935-openclaw-skills-ai-agents-frameworks", "kind": "skill", "name": "ai_agents_frameworks", "description": "Reason about multi-step agent systems, tool use, orchestration, and control loops using concepts from LangGraph, CrewAI, tool-calling examples, and agent courses. Use when building AI agents that plan, call tools, manage state, or coordinate subtasks.", "publisher": "Wang200935", "version": "1.0.0", "capabilities": { "domains": [ "general" ], "tags": [ "skill-md", "github" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "Reason about multi-step agent systems, tool use, orchestration, and control loops using concepts from LangGraph, CrewAI, tool-calling examples, and agent courses. Use when building AI agents that plan, call tools, manage state, or coordinate subtasks." ], "primary": false, "metadata": { "source": { "provider": "github", "repository": "https://github.com/Wang200935/openclaw-skills", "path": "misc/_merge_backups/auto_20260315_070002/misc/ai/ai-agents-frameworks/SKILL.md", "ref": "8f2b5b95360219519f8692b3056a02a12aa76487", "url": "https://github.com/Wang200935/openclaw-skills/blob/8f2b5b95360219519f8692b3056a02a12aa76487/misc/_merge_backups/auto_20260315_070002/misc/ai/ai-agents-frameworks/SKILL.md", "key": "Wang200935/openclaw-skills/misc/_merge_backups/auto_20260315_070002/misc/ai/ai-agents-frameworks/SKILL.md" } }, "instructions": "# Agent Framework Patterns\n\nUse this skill for agentic systems.\n\n## Core rules\n\n- Keep control loops explicit.\n- Make tool boundaries auditable.\n- Separate planner state from executor state.\n- Add guardrails around destructive actions.\n- Prefer deterministic repair flows over free-form autonomy for ops tasks.\n\n## Common pattern\n\nobserve -> diagnose -> decide -> act -> verify\n\n## For ops agents\n\n- Log every action.\n- Require confirmation for dangerous steps unless already authorized.\n- Make retries bounded and visible.\n\n\n\n---\nMERGED_FROM_BACKUP: C:\\Users\\wang\\.openclaw\\skills\\ai\\ai-agents-frame", "cost": { "context_tokens": 360 } }

ai_agents_frameworks

About

Agent Framework Patterns

Core rules

Common pattern

For ops agents

MERGED_FROM_BACKUP: C:\Users\wang.openclaw\skills\ai\ai-agents-frameworks\SKILL.md

name: ai_agents_frameworks description: Reason about multi-step agent systems, tool use, orchestration, and control loops using concepts from LangGraph, CrewAI, tool-calling examples, and agent courses. Use when building AI agents that plan, call tools, manage state, or coordinate subtasks.

Agent Framework Patterns

Core rules

Common pattern

For ops agents

Use it

Manifest

Reviews