Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/JinLee794/agent-framework-skillsnpx agentmods add skills/jinlee794/agent-framework-skills/maf-multi-agent-workflowsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/jinlee794/agent-framework-skills/maf-multi-agent-workflows)<a href="https://agentmods.dev/skills/jinlee794/agent-framework-skills/maf-multi-agent-workflows"><img src="https://agentmods.dev/badge/skills/jinlee794/agent-framework-skills/maf-multi-agent-workflows.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00102 | $0.03523 |
| Opus 5 | $0.00051 | $0.01761 |
| Sonnet 5 | $0.00020 | $0.00705 |
| Haiku 4.5 | $0.00010 | $0.00352 |
Grade A, and why
maf-multi-agent-workflows scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 321 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Multi-Agent Structure & Workflow Orchestration
Single-agent construction — clients, @tool, skills, sessions, middleware, retrieval — is
owned by maf-foundry-agent. The agent document schema is owned
by maf-agent-config. This skill starts at the second agent.
| Task | Reference |
|---|---|
| Pick and wire one of the five built-in orchestrations | references/orchestration-patterns.md |
| Hand-build a graph: executors, edges, events, checkpoints, HITL | references/graph-workflows.md |
First, do not split
Every added participant is another model round trip. Behind a live call that is dead air, and the caller hears it. Split only when one of these is true:
- Two instruction sets genuinely contradict each other (a refund policy and a sales pitch).
- The tool surface has grown past reliable selection — roughly ten tools in one list.
- A step must run without the caller's conversation in context (summarise, classify, redact).
- A stage needs a different approval boundary or a narrower data reach than the rest.
Otherwise add a tool, not an agent. If you only need one delegated call and no shared
conversation, use agent.as_tool(...) and stop — that is
maf-foundry-agent, not a workflow.
Layout
Extends the canonical tree in maf-voice-agent; this skill owns the multi-agent parts of it.
config/
agents/
triage.yaml # kind: Prompt — one flat file per participant
refunds.yaml
orders.yaml
workflows/
support.yaml # kind: Workflow — references the agent files above
voice/
support.yaml # mounts: {workflow: support} — the runnable runtime
src/<package>/
agents/
__init__.py # re-exports build_*_agent factories, nothing else
triage.py # only when a participant needs code-only wiring
workflows/
__init__.py # re-exports build_*_workflow factories
support.py # one module per workflow: participants in, Workflow out
executors/
redact.py # non-agent nodes only
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 321 lines · 102 tokens per session scan A c1a6a9ca1deb
maf-multi-agent-workflows is a skill published in the GitHub repository JinLee794/agent-framework-skills (2 stars, last pushed 1mo ago), licensed MIT. It adds 102 tokens to every session and 3,523 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
claude-on-foundry
End-to-end assistant skill for the Claude on Foundry Starter Kit (Azure-Samples/claude). Walks customers through deploying, verifying, modifying, debugging, and tearing down a Claude model deployment on Microsoft Foundry using either the Bicep or Terraform IaC variant in this repo, with one-command guidance via azd…
aspire-deployment
WORKFLOW SKILL — Deploy Aspire apps from AppHost models to Docker Compose, Kubernetes, Azure, or AWS. WHEN: "deploy Aspire app", "publish Aspire artifacts", "deploy to Azure Container Apps", "generate Kubernetes artifacts", "tear down Aspire deployment". INVOKES: aspire CLI, Aspire docs, target cloud/container CLIs.…
entra-a2a-mcp-obo
Best practices, tips, and gotchas for Entra ID, Entra Agent ID, A2A protocol, MCP protocol, and agentgateway-based OBO token exchange — distilled from loop-runtime's entraagentid.md docs. Use when setting up, wiring, or debugging Entra app registrations, Entra Agent ID blueprints, OBO chains, A2A/MCP auth, app-code…
aspire-monitoring
ANALYSIS SKILL - Observe Aspire apps: logs, traces, metrics, resource state, telemetry export, browser telemetry, and the standalone dashboard. Routes between local Aspire CLI, AKS workload diagnostics, and deployed Azure resource health. USE FOR: aspire logs, aspire otel logs, aspire otel traces, aspire otel spans…
harness-engineering
Adopt repository-level harness engineering for coding agents. Use when a user wants to prevent repeated AI coding-agent mistakes by turning failures into durable instructions, drift checks, regression tests, failure memory, and adoption reports tailored to the target repository.
get-api-docs
Use this skill when you need documentation for a third-party library, SDK, or API before writing code that uses it — for example, "use the OpenAI API", "call the Stripe API", "use the Anthropic SDK", "query Pinecone", or any time the user asks you to write code against an external service and you need current API…