"delegate", "send to agent", "dispatch"?
send_task / send_task_streaming
← Skill registry
Agent infrastructurePublic skills
A2A Agent Orchestration
Multi-agent coordination — discover capabilities, dispatch tasks, stream results, and manage remote agent lifecycles via the A2A protocol.
What this skill teaches
A reusable playbook for a specific kind of work.
An MCP server tells an agent which actions are available. A skill adds the judgment around those actions: how to recognize the job, which sequence to follow, what to avoid, and how to decide that the result is complete.
Orchestrate remote AI agents via A2A protocol — discover agent cards, dispatch tasks, stream results, cancel tasks, and manage push notifications. Use when delegating work to remote agents, checking agent capabilities, coordinating multi-agent workflows, or monitoring task progress.
Architecture
How A2A Agent Orchestration guides an ADK-Rust agent.
The skill stays readable and portable because it contains instructions rather than service credentials or business data. ADK-Rust supplies it to the agent, the agent chooses from its reviewed tool boundary, and the connected MCP server performs the authenticated operation.
01
User request
The agent receives a goal expressed in ordinary language.
02
A2A Agent Orchestration skill
Matches intent, supplies the decision guide, and narrows the tool boundary.
03
ADK-Rust agent
Plans the workflow and streams each meaningful step through the runtime.
04
mcp-a2a
Executes authenticated operations against the system that owns the capability.
05
Verified result
The skill's completion rules shape the evidence returned to the user.
Decision guide
How the agent turns a request into the right action.
These routes come directly from the skill instructions. They help the model recognize intent and select a focused tool or workflow instead of improvising across the entire capability surface.
"what agents", "capabilities", "who can"?
list_agents / fetch_agent_card
"task status", "progress", "check on"?
get_task / list_tasks
"cancel", "stop task"?
cancel_task
"subscribe", "notifications"?
subscribe_events / manage_push_notifications
Proven workflows
Repeatable sequences for useful outcomes.
A workflow joins several tool calls into a task the user actually recognizes. The skill explains the sequence and the intended result while ADK-Rust streams the agent's progress through the shared runtime.
013 calls
Discover & Dispatch
Find agent → verify → send task
021 calls
Stream Results
Real-time partial outputs
031-2 calls
Monitor
Track task to completion
041 calls
Cancel
Abort stuck tasks
Tool boundary
The skill may use 10 documented tools.
This allowlist is declared by the skill. It keeps the agent focused on the actions needed for this job while mcp-a2a retains responsibility for authentication, validation, and the connected system.
fetch_agent_card
validate_agent_card
list_agents
send_task
send_task_streaming
get_task
list_tasks
cancel_task
subscribe_events
manage_push_notifications
Working rules
What the agent should do.
- Validate agent card before dispatching sensitive data
- Set explicit timeouts on all tasks
- Track every task to completion (don't fire-and-forget)
- Use streaming for long-running tasks (better UX)
Operating boundaries
What the agent should avoid.
- Don't send sensitive data to unvalidated agents
- Don't leave tasks running indefinitely without timeout
- Don't dispatch without checking agent is available
Install and connect
Add the skill beside the capability it expects.
Install the repository where your ADK-Rust skill loader can discover it, connect mcp-a2a, and confirm the declared tools are available before asking the agent to use the workflow.
Install the skill
git clone https://github.com/zavora-ai/skill-a2a-agent-orchestration.git \
~/.skills/skills/a2a-agent-orchestrationADK-Rust loading shape
let skills = SkillLoader::from_dir("~/.skills/skills").await?;
let skill = skills.load("a2a-agent-orchestration").await?;
let agent = LlmAgentBuilder::new("agent")
.instruction(skill.instructions())
.tools(skill.allowed_tools(toolset)?)
.build()?;The repository's compatibility statement: Requires mcp-a2a server connected.
Official documentation
Read the complete skill package.
The repository remains authoritative for its exact instructions, examples, helper scripts, assets, MCP requirements, license, and later updates.
Keep exploring
More in agent infrastructure.
ACP Workspace ManagementCoding agent workspace control — start sessions, delegate tasks to coding agents, review proposed patches, manage permissions, and publish agents with governance gates.9 allowed toolsArtifact LifecycleGoverned artifact management — versioned storage, provenance tracking, integrity verification, derivation chains, and export packaging for build outputs and documents.10 allowed toolsCloud InfrastructureCloud operations for AI agents — deploy services, scale replicas, manage DNS, rotate secrets, promote between environments, monitor costs, and rollback on failure.24 allowed tools
Source record
Repository metadata for this skill entry.
- License
- Apache-2.0
- Allowed tools
- 10
- References
- 3
- Revision
- af1f68ba5644
- Updated
- May 31, 2026