abhisheksharma-17/ms-agent-framework
skills/ms-agent-framework/SKILL.md
Expert skill for building AI agent systems with Microsoft Agent Framework (Python). MANDATORY TRIGGERS: Microsoft Agent Framework, agent-framework, Azure AI agents, Azure OpenAI agents. Also trigger when the user is building production AI agents, multi-agent orchestrations, agentic workflows on Azure, or discussing concepts like tools, semantic kernel, middleware, or memory in Python. When in doubt, use it.
tools
Updated Mar 27, 2026
$ install --global
skillsauthnpx skillsauth add abhisheksharma-17/skills-graph ms-agent-frameworkInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Mar 28, 2026, 2:34 AM36.0s69 files scanned
SKILL.md
- name:
- ms-agent-framework
- description:
- Expert skill for building AI agent systems with Microsoft Agent Framework (Python). MANDATORY TRIGGERS: Microsoft Agent Framework, agent-framework, Azure AI agents, Azure OpenAI agents. Also trigger when the user is building production AI agents, multi-agent orchestrations, agentic workflows on Azure, or discussing concepts like tools, semantic kernel, middleware, or memory in Python. When in doubt, use it.
- license:
- MIT
- version:
- 1.0.0
- author:
- Abhishek Sharma
- tags:
- ["agent framework", "python", "azure", "multi-agent"]
Microsoft Agent Framework — Python Expert Skill
Version: 1.0.0b260130 (Public Preview) | Python only | Repo: https://github.com/microsoft/agent-framework
Reference Files
| Reference | File | Read When |
|-----------|------|-----------|
| Framework Overview | references/00-framework-overview.md | "what is this framework", "critical rules", "setup", "architecture", "common errors" |
| Getting Started | references/01-getting-started.md | "create a bot", "AI assistant", "hello world" |
| Running Agents | references/02-running-agents.md | "run agents", "streaming vs non-streaming", "ResponseStream" |
| Structured Output | references/03-structured-output.md | "JSON output", "typed response", "Pydantic" |
| Tools (Function) | references/04-tools-function.md | "call a function", "agent uses tools", "@tool decorator" |
| Tools (Hosted) | references/05-tools-hosted.md | "search the web", "MCP servers", "code interpreter" |
| RAG | references/06-rag.md | "search docs", "knowledge base" |
| Sessions & Memory | references/07-sessions.md | "remember things", "conversation history" |
| Memory Providers | references/08-memory.md | context providers, custom memory providers |
| Middleware | references/09-middleware.md | "intercept", "filter", "transform output" |
| Model Providers | references/10-providers.md | Azure OpenAI, OpenAI, Anthropic, Ollama, GitHub |
| Workflows | references/11-workflows-core.md | "pipeline", "execution graph", handlers, edges, events |
| Multi-Agent | references/12a-orchestration-sequential.md | "multiple agents", "coordination", sequential, parallel, handoff |
| Deployment | references/13-deployment.md | "Azure Functions", "host serverless", "A2A", "OpenAI compatible" |
| Declarative | references/14-declarative.md | "YAML workflow", "config-based agent" |
| Observability | references/15-observability.md | "debug", "trace", "OpenTelemetry" |
| Multimodal | references/16-multimodal.md | "vision", "analyze images" |
| Custom Agents | references/17-custom-agents.md | "own agent", "extend framework" |
| Security & Purview | references/19-security.md | "guardrails", "content safety", compliance, audit |
| M365 Integration | references/21-m365-integration.md | "Copilot agent", "Teams", "Graph API" |
| Design Patterns | references/22-design-patterns-core.md | "pattern for reflection / debate / supervisor / pipeline" |
| Advanced Patterns | references/23-design-patterns-advanced.md | circuit breakers, load balancing, batching |
Installation
pip install agent-framework --pre
az login # For Azure authentication
Version Tracking
- Skill version: 1.0.0 | Snapshot: 2026-02-21
- Version metadata:
VERSION.json - Update checker:
python scripts/check-updates.py - Changelog:
CHANGELOG.md
Related Skills
abhisheksharma-17/weaviate
tools
VerifiedTrustedCommunity
Open-source AI-native vector database for semantic search, hybrid search, RAG, and agent-driven workflows with multi-tenancy and model provider integrations. MANDATORY TRIGGERS: weaviate, Weaviate, weaviate-client, near_text, near_vector, hybrid search vector database, WeaviateClient, weaviate.connect, collections.create, query.near_text, query.hybrid, query.bm25. Also trigger when user wants to build semantic search, store vector embeddings, implement RAG pipelines, combine keyword and vector search, set up multi-tenant vector storage, use named vectors, or integrate vector DB with LLM providers. When in doubt about whether to use this skill for vector database or semantic search tasks, use it.
1SKILL.mdUpdated Jun 4, 2026
abhisheksharma-17/playwright
tools
VerifiedTrustedCommunity
End-to-end testing and browser automation framework by Microsoft — locators, assertions, fixtures, network mocking, visual testing, tracing, codegen, API testing, and CI/CD integration. MANDATORY TRIGGERS: playwright, Playwright, @playwright/test, page.goto, page.click, page.locator, getByRole, getByText, getByTestId, expect(page), toHaveScreenshot, playwright.config, npx playwright. Also trigger when user wants to write E2E tests, automate browser interactions, test across Chromium/Firefox/WebKit, mock network requests in tests, do visual regression testing, generate tests with codegen, or set up CI test pipelines. When in doubt about whether to use this skill for browser testing or E2E automation tasks, use it.
1SKILL.mdUpdated Jun 4, 2026
abhisheksharma-17/openai-agents-sdk
tools
VerifiedTrustedCommunity
OpenAI Agents SDK for building multi-agent workflows with tools, handoffs, guardrails, streaming, MCP, sessions, and tracing. MANDATORY TRIGGERS: openai-agents, openai agents sdk, openai-agents-python, openai agents, Runner.run, function_tool, handoff, guardrail, MCPServerStdio, MCPServerStreamableHttp, HostedMCPTool, RunContextWrapper, AgentHooks. Also trigger when user wants to build multi-agent systems with OpenAI models, create agent orchestration with handoffs, add guardrails to LLM applications, integrate MCP servers with agents, implement streaming agent responses, or use OpenAI's official agent framework. When in doubt about whether to use this skill for OpenAI agent tasks, use it.
1SKILL.mdUpdated Jun 4, 2026
abhisheksharma-17/haystack
tools
VerifiedTrustedCommunity
AI orchestration framework for building production-ready RAG applications, autonomous agents, and multimodal search systems. MANDATORY TRIGGERS: haystack, deepset, haystack-ai, haystack pipeline, haystack agent. Also trigger when the user wants to build RAG pipelines with modular components, create tool-calling agents with Haystack, orchestrate retrieval-augmented generation, build semantic search systems, or evaluate LLM pipelines. When in doubt about whether to use this skill for RAG orchestration or AI pipeline tasks, use it.
1SKILL.mdUpdated Jun 4, 2026