Swarm Orchestration

This skill provides the orchestration framework for running predefined multi-agent teams. The swarm-coordinator agent uses these patterns to spawn, sequence, and aggregate agent workflows.

When to Use

•/blueprint-dev:bp:team-* commands — predefined team swarms
•/blueprint-dev:bp:lfg — full pipeline orchestration
•When the swarm-coordinator needs to understand team composition

Team Catalog

See references/team-catalog.md for predefined teams and their compositions.

Orchestration Patterns

Sequential

Agents run one after another, each using the previous agent's output.

code

A → B → C

Use when: Agent B needs Agent A's output to function.

Parallel

All agents run simultaneously on the same input.

code

[A, B, C] → merge

Use when: Agents are independent and evaluate different aspects.

Sequential-then-Parallel

One agent produces output, then multiple agents review it in parallel.

code

A → [B, C, D] → merge

Use when: One agent creates, multiple agents evaluate.

Pipeline

Teams run sequentially, each team's output feeds the next.

code

team-design → team-architecture → team-review

Use when: Full workflow execution with approval gates between phases.

Result Aggregation Rules

•Deduplicate: If multiple agents flag the same issue, keep one with the highest priority
•Standardize priorities: Ensure P1/P2/P3 meanings are consistent across agents
•Attribute: Every finding shows which agent produced it
•Conflicts: Flag when agents disagree for user resolution

Approval Gates

For the full pipeline (/blueprint-dev:bp:lfg), insert user approval gates between phases:

•After planning: "Approve plan before designing?"
•After design: "Approve variants before architecture review?"
•After architecture: "Approve architecture before building?"
•After review: "Approve for shipping?"