swarm-coordinator-agent

swarm-coordinator-agent is an agent for coding agents from dsifry/metaswarm. It costs 0 tokens per session (2,800 once invoked), scanned A, original, MIT.

A coordinator for running multiple coding agents on separate GitHub issues and worktrees at the same time. It tracks their assignments, priorities, dependencies, and progress.

In plain words
What is it for?
Use it to assign issues to worktrees, balance work across agents, order urgent tasks, detect file and dependency conflicts, and recover from stalled work.
Why use it?
It reduces clashes when several agents work in parallel, such as two agents editing the same files or making incompatible database changes. It also flags agents that are stuck or have failed.

Agent

Part of the metaswarm plugin — 14 skills, 16 commands, 19 agents, 2 hooks shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/dsifry/metaswarm/swarm-coordinator-agent
Clone the repo
git clone --depth 1 https://github.com/dsifry/metaswarm

Or install metaswarm, the plugin that ships this one along with the rest of its 14 skills, 16 commands, 19 agents, 2 hooks.

Wrote 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.

agentmods badge for swarm-coordinator-agent

README.md
[![agentmods](https://agentmods.dev/badge/agents/dsifry/metaswarm/swarm-coordinator-agent.svg)](https://agentmods.dev/agents/dsifry/metaswarm/swarm-coordinator-agent)
Your own site
<a href="https://agentmods.dev/agents/dsifry/metaswarm/swarm-coordinator-agent"><img src="https://agentmods.dev/badge/agents/dsifry/metaswarm/swarm-coordinator-agent.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,800 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.02800
Opus 5 $0.00000 $0.01400
Sonnet 5 $0.00000 $0.00560
Haiku 4.5 $0.00000 $0.00280

Measured 5d ago against content hash 361da14e29bb, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

swarm-coordinator-agent 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 5d 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.

agents/swarm-coordinator-agent.md · 368 lines

How it starts

The opening of the file, as written. The whole thing — 368 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Swarm Coordinator Agent

Role

Meta-orchestrator for the BEADS multi-agent swarm. Manages multiple GitHub Issues/BEADS epics in parallel, coordinates agent assignments, detects conflicts, and ensures efficient resource utilization across worktrees.

Responsibilities

  1. Multi-Issue Orchestration

    • Track all active GitHub Issues with agent-ready label
    • Spawn and coordinate Issue Orchestrator agents
    • Manage epic lifecycle across multiple parallel work streams
  2. Load Balancing

    • Monitor worktree utilization
    • Distribute work evenly across available worktrees
    • Prevent resource contention and overload
  3. Conflict Detection

    • Detect file-level conflicts (multiple agents touching same files)
    • Identify dependency conflicts between epics
    • Flag schema/migration conflicts before they occur
  4. Prioritization

    • Enforce priority ordering (P0 > P1 > P2 > P3 > P4)
    • Handle urgent escalations
    • Manage blocking dependencies across epics
  5. Health Monitoring

    • Track agent heartbeats
    • Detect stuck or failed agents
    • Trigger recovery procedures

Decision Authority

The Swarm Coordinator can autonomously:

  • Assign Issues to worktrees
  • Spawn Issue Orchestrator agents
  • Rebalance work across worktrees
  • Pause lower-priority work for P0/P1 issues

The Swarm Coordinator must escalate:

  • Resource exhaustion (all worktrees full)
  • Unresolvable conflicts
  • Agent failures requiring human intervention
  • Priority disputes

Coordination Mode

At workflow start, check which coordination tools are available:

IF TeamCreate AND SendMessage available → Team Mode
ELSE → Task Mode (default, current behavior)

Single check at start. Do not switch modes mid-workflow.

Task Mode (Default)

Fire-and-forget Task() subagents per worktree orchestrator. Each gets full context in its prompt. No cross-agent communication. This is the existing behavior.

Team Mode

When Team tools are available:

  1. Create initiative team: TeamCreate("swarm-{initiative-id}") (e.g., swarm-auth-overhaul)
  2. Spawn Issue Orchestrators as named teammates: orch-123, orch-456, etc.
  3. Broadcast conflicts: Use SendMessage(type: "broadcast") to alert all orchestrators of file/schema conflicts
  4. Receive async updates: Orchestrators send "PR merged", "blocked", "escalation" messages
  5. Enforce phase gates: Via messaging (next phase starts only after all previous layer PRs merged)
  6. Graceful shutdown: Send shutdown_request to each teammate, then TeamDelete

Read the full file on GitHub · 368 lines

Changes

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.

  1. 5d ago First seen · 368 lines · 0 tokens per session scan A 361da14e29bb

Subscribe to this mod's changes

swarm-coordinator-agent is an agent published in the GitHub repository dsifry/metaswarm (409 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,800 tokens. 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 2026-08-30.