Back to search
agentic-ai-systems
SkillGitHub
Agentic AI explained with chickens π every pattern a runnable, CI-checked file.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Checked today
Badge for your README
Markdown
[](/listings/thibautmelen--agentic-ai-systems)HTML
<a href="/listings/thibautmelen--agentic-ai-systems"><img src="/api/listings/thibautmelen--agentic-ai-systems/badge.svg" alt="findsafeskills safety rating" /></a>Install
i
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md Β· README Β· 321 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
- <div align="center">
- # Agentic AI Systems π
- **Agentic systems explained with chickens β because AI orchestration doesn't have to be scary**
- *Main Agent spawns Subagents like a hen with her chicks ππͺΊπ¦*
- [](https://github.com/ThibautMelen/agentic-ai-systems/actions/workflows/check.yml)
- [](https://docs.anthropic.com/en/docs/claude-code)
- [](https://www.anthropic.com/research/building-effective-agents)
- [](https://github.com/SuperNovae-studio)
- [](https://github.com/hesreallyhim/awesome-claude-code)
- [](LICENSE)
- [](CONTRIBUTING.md)
- [](https://mermaid.js.org/)
- [π¦ Foundations](foundations/) β’ [βοΈ Workflows](workflows/) β’ [π Autonomous](agents/) β’ [π As Code](patterns-as-code/) β’ [π οΈ Implementation](implementation/) β’ [πΊοΈ Guides](guides/)
- </div>
- ---
- ## Overview
- ```mermaid
- %%{init: {'theme': 'base', 'themeVariables': { 'primaryColor': '#8b5cf6', 'primaryTextColor': '#fff', 'primaryBorderColor': '#7c3aed', 'lineColor': '#a78bfa', 'secondaryColor': '#ec4899', 'tertiaryColor': '#6366f1', 'noteTextColor': '#fff', 'noteBkgColor': '#8b5cf6', 'textColor': '#fff' }}}%%
- mindmap
- root((π Agentic Systems))
- π¦ Foundations
- Augmented LLM
- What Changed 2026
- βοΈ Workflows
- ποΈ Baseline
- βοΈ Chaining
- π¦ Routing
- π€οΈ Parallel
- π¦ Orchestrator
- π©» Evaluator
- π Autonomous Agent
- The Alternative
- Multi-Window
- π οΈ Implementation
- π¦ Subagent
- 𦴠Command
- π Skill
- πͺ Hook
- π Patterns as Code
- Check before
- Trace after
- ```
- ---
- > [!TIP]
- > **New in the 2026 edition** β [ποΈ What Changed 2025-2026](foundations/what-changed-2026.md) (the dated map: context engineering, the multi-agent debate, the AGENTS.md Β· SKILL.md Β· MCP stack, harness engineering) and [π Patterns as Code](patterns-as-code/) (every pattern as a runnable file β checked in CI on every push, no API key needed).
- <p align="center">
- <a href="patterns-as-code/"><img src="https://raw.githubusercontent.com/supernovae-st/nika/main/media/gifs/chat-to-workflow.optimized.gif" alt="A repeated chat prompt becomes a workflow file; checked before a token is spent, then a real local run β the patterns in this repo, as runnable files" width="850"></a>
- </p>
- <p align="center"><sub>β¬οΈ real capture: a repeated prompt graduates into a checked, runnable file β <a href="patterns-as-code/">every pattern in this repo works this way</a> Β· and <a href="patterns-as-code/real-world/">three of them run a real campaign daily</a> π°οΈ</sub></p>
- ## πΊοΈ Navigation
- <table>
- <tr>
- <td width="50%" valign="top">
- ### π¦ [Foundations](foundations/)
- *The building block for everything*
- | | |
- |---|---|
- | [π¦ Augmented LLM](foundations/augmented-llm.md) | LLM + Retrieval + Tools + Memory |
- | [ποΈ What Changed 2025-2026](foundations/what-changed-2026.md) | Context engineering Β· skills/MCP stack Β· harness era |
- ---
- ### βοΈ [Workflows](workflows/)
- *Predefined orchestration β code controls the flow*
- | # | Workflow | Use When |
- |:-:|----------|----------|
- | 0 | [ποΈ Baseline](workflows/00-baseline.md) | Simple, 1-step task |
- | 1 | [βοΈ Prompt Chaining](workflows/01-prompt-chaining.md) | Sequential steps |
- | 2 | [π¦ Routing](workflows/02-routing.md) | Classify & dispatch |
- | 3 | [π€οΈ Parallelization](workflows/03-parallelization.md) | Independent tasks |
- | 4 | [π¦ Orchestrator](workflows/04-orchestrator-workers.md) | Expert delegation |
- | 5 | [π©» Evaluator](workflows/05-evaluator-optimizer.md) | Quality iteration |
- </td>
- <td width="50%" valign="top">
- ### π [Autonomous Agent](agents/)
- *Dynamic autonomy β LLM controls the flow*
- | Pattern | Use When |
- |---------|----------|
- | [π Autonomous Agent](agents/autonomous.md) | Open-ended problems |
- | [π₯οΈ Multi-Window](agents/multi-window.md) | Cross-session state (variant) |
- ---
- ### π οΈ [Implementation](implementation/)
- *Claude Code components & architecture*
- | Component | Location |
- |-----------|----------|
- | [π¦ Subagent](implementation/components/subagent.md) | `.claude/agents/*.md` |
- | [𦴠Command](implementation/components/slash-command.md) | `.claude/commands/*.md` |
- | [π Skill](implementation/components/skill.md) | `.claude/skills/*/SKILL.md` |
- | [πͺ Hook](implementation/components/hook.md) | `.claude/settings.json` |
- ---
- ### πΊοΈ [Guides](guides/) & [π Reference](reference/)
- | Resource | Description |
- |----------|-------------|
- | [Selection Guide](guides/README.md) | Choose the right pattern |
- | [π Patterns as Code](patterns-as-code/) | Every pattern as a runnable, checkable file |
- | [Use Cases](guides/use-cases/) | 6 validated examples |
- | [Glossary](reference/glossary.md) | A-Z definitions |
- | [Visual Standards](reference/visual-standards.md) | Colors & emojis |
- </td>
- </tr>
- </table>
- ---
- ## Quick Decision
- ```mermaid
- %%{init: {'theme': 'base', 'themeVariables': {'lineColor': '#64748b'}}}%%
- flowchart LR
- START((π― Task)) --> DEST{Destructive?}
- DEST -->|Yes| WIZ[π§ Wizard]
- DEST -->|No| COMP{Complex?}
- COMP -->|No| BASE[ποΈ Baseline]
- COMP -->|Yes| PRED{Predictable<br/>steps?}
- PRED -->|Yes| WORK{Need<br/>specialists?}
- PRED -->|No| AGENT[π Autonomous]
- WORK -->|No| CHAIN[βοΈ Chain]
- WORK -->|Yes| ORCH[π¦ Orchestrator]
- classDef default fill:#f8fafc,stroke:#64748b,stroke-width:1px,color:#1e293b
- classDef decision fill:#fef3c7,stroke:#f59e0b,stroke-width:2px,color:#92400e
- classDef baseline fill:#64748b,stroke:#475569,stroke-width:2px,color:#ffffff
- classDef wizard fill:#14b8a6,stroke:#0d9488,stroke-width:2px,color:#ffffff
- classDef workflow fill:#8b5cf6,stroke:#7c3aed,stroke-width:2px,color:#ffffff
- classDef agent fill:#ec4899,stroke:#db2777,stroke-width:2px,color:#ffffff
- START:::decision
- DEST:::decision
- COMP:::decision
- PRED:::decision
- WORK:::decision
- BASE:::baseline
- WIZ:::wizard
- CHAIN:::workflow
- ORCH:::workflow
- AGENT:::agent
- ```
- | Situation | β Use |
- |-----------|-------|
- | Simple task (1 step) | [ποΈ Baseline](workflows/00-baseline.md) |
- | Sequential (2-4 steps) | [βοΈ Prompt Chaining](workflows/01-prompt-chaining.md) |
- | Categorize inputs | [π¦ Routing](workflows/02-routing.md) |
- | Independent subtasks | [π€οΈ Parallelization](workflows/03-parallelization.md) |
- | Multiple specialists | [π¦ Orchestrator-Workers](workflows/04-orchestrator-workers.md) |
- | Quality iteration | [π©» Evaluator-Optimizer](workflows/05-evaluator-optimizer.md) |
- | Open-ended / unknown steps | [π Autonomous Agent](agents/autonomous.md) |
- | Destructive operations | [π§ Wizard](workflows/01-prompt-chaining.md#advanced-wizard-pattern) |
- | Long-running (>10 min) | [π₯οΈ Multi-Window Context](agents/multi-window.md) |
- ---
- ## Anthropic Taxonomy
- ```mermaid
- %%{init: {'theme': 'base', 'themeVariables': {'lineColor': '#64748b'}}}%%
- flowchart LR
- subgraph WORKFLOWS["βοΈ WORKFLOWS"]
- direction TB
- W1[ποΈ Baseline]
- W2[βοΈ Prompt Chaining]
- W3[π¦ Routing]
- W4[π€οΈ Parallelization]
- W5[π¦ Orchestrator]
- W6[π©» Evaluator]
- end
- subgraph AGENTS["π AUTONOMOUS AGENT"]
- direction TB
- A1[π The Alternative]
- A2[π₯οΈ Multi-Window variant]
- end
- CODE[π Code controls] --> WORKFLOWS
- WORKFLOWS --> LLM[π§ LLM controls]
- LLM --> AGENTS
- classDef workflowBox fill:#ede9fe,stroke:#8b5cf6,stroke-width:2px,color:#5b21b6
- classDef agentBox fill:#fce7f3,stroke:#ec4899,stroke-width:2px,color:#9d174d
- classDef control fill:#f1f5f9,stroke:#64748b,stroke-width:1px,color:#475569
- WORKFLOWS:::workflowBox
- AGENTS:::agentBox
- CODE:::control
- LLM:::control
- ```
- > **Key distinction:** Workflows have predefined paths (code controls). Agents decide their own path (LLM controls).
- ---
- ## Critical Rule
- ```mermaid
- %%{init: {'theme': 'base', 'themeVariables': {'lineColor': '#64748b'}}}%%
- flowchart LR
- U1[πββοΈ User] -->|request| MA[π Main Agent]
- MA -->|πͺΊ spawn| SA1[π¦ Subagent]
- MA -->|πͺΊ spawn| SA2[π¦ Subagent]
- SA1 -->|result| MA
- SA2 -->|result| MA
- MA -->|response| U2[πββοΈ User]
- SA1 x--x|"β CANNOT spawn"| SA3[π¦]
- classDef user fill:#6366f1,stroke:#4f46e5,stroke-width:2px,color:#ffffff
- classDef main fill:#f59e0b,stroke:#d97706,stroke-width:2px,color:#ffffff
- classDef sub fill:#ec4899,stroke:#db2777,stroke-width:2px,color:#ffffff
- classDef blocked fill:#ef4444,stroke:#dc2626,stroke-width:2px,color:#ffffff,stroke-dasharray: 5 5
- U1:::user
- U2:::user
- MA:::main
- SA1:::sub
- SA2:::sub
- SA3:::blocked
- ```
- > **π¦ Subagents cannot spawn other π¦ subagents.** All delegation flows through π Main Agent.
- ---
- ## Repository Structure
- ```
- .
- βββ README.md # π You are here
- β
- βββ foundations/ # π¦ Core concepts
- β βββ augmented-llm.md
- β
- βββ workflows/ # βοΈ Predefined orchestration
- β βββ 00-baseline.md
- β βββ 01-prompt-chaining.md
- β βββ 02-routing.md
- β βββ 03-parallelization.md
- β βββ 04-orchestrator-workers.md
- β βββ 05-evaluator-optimizer.md
- β
- βββ agents/ # π Autonomous Agent (the alternative)
- β βββ autonomous.md # The pattern
- β βββ multi-window.md # Variant
- β
- βββ implementation/ # π οΈ Claude Code specifics
- β βββ components/ # π¦π¦΄ππͺ
- β βββ architecture/ # 5-layer system
- β
- βββ guides/ # πΊοΈ Selection & use cases
- β βββ use-cases/ # 6 validated examples
- β
- βββ reference/ # π Glossary, standards
- ```
- ---
- ## References
- | Resource | Link |
- |----------|------|
- | Building Effective Agents | [anthropic.com/engineering](https://www.anthropic.com/engineering/building-effective-agents) |
- | Claude Code Docs | [docs.anthropic.com](https://docs.anthropic.com/en/docs/claude-code) |
- | Agent SDK | [docs.anthropic.com/agent-sdk](https://docs.anthropic.com/docs/en/agent-sdk) |
- | Anthropic Cookbook | [github.com/anthropics](https://github.com/anthropics/anthropic-cookbook) |
- | Agent Skills spec | [agentskills.io](https://agentskills.io/specification) |
- | Agentic AI Foundation (MCP Β· AGENTS.md Β· goose) | [aaif.io](https://aaif.io/) |
- | Patterns as runnable files (this repo) | [patterns-as-code/](patterns-as-code/) |
- ---
- ## Contributing
- Contributions welcome! See [CONTRIBUTING.md](CONTRIBUTING.md).
- **Requirements:** Official sources β’ Code examples β’ Mermaid diagrams β’ Established format
- ---
- <div align="center">
- <sub>Built with Claude Code | Based on Anthropic documentation | 2025-2026</sub><br/>
- <sub>Independent community resource β not affiliated with Anthropic</sub>
- <br/>
- <a href="https://github.com/ThibautMelen">
- <img src="https://avatars.githubusercontent.com/u/20891897?s=200&v=4" alt="ThibautMelen" width="32"/>
- </a>
- β€οΈ
- <a href="https://github.com/SuperNovae-studio">
- <img src="https://avatars.githubusercontent.com/u/33066282?s=200&v=4" alt="SuperNovae Studio" width="32"/>
- </a>
- π΄ββ οΈ
- </div>
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If itβs helped, pass it on.