Borrowing it
Nothing to install: this file belongs to saeed-vayghan/gemini-agent-skills. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/saeed-vayghan/gemini-agent-skills/master/.gemini/skills/multi-agent-coordinator/SKILL.mdgit clone --depth 1 https://github.com/saeed-vayghan/gemini-agent-skillsWrote 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/saeed-vayghan/gemini-agent-skills/multi-agent-coordinator)<a href="https://agentmods.dev/skills/saeed-vayghan/gemini-agent-skills/multi-agent-coordinator"><img src="https://agentmods.dev/badge/skills/saeed-vayghan/gemini-agent-skills/multi-agent-coordinator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/saeed-vayghan/gemini-agent-skills/multi-agent-coordinator"><img src="https://agentmods.dev/badge/skills/saeed-vayghan/gemini-agent-skills/multi-agent-coordinator.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00044 | $0.01217 |
| Opus 5.5 | $0.00018 | $0.00487 |
| Sonnet 5.5 | $0.00009 | $0.00243 |
| Haiku 4.5 | $0.00004 | $0.00122 |
Grade A, and why
multi-agent-coordinator 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 1mo 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 — 269 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior multi-agent coordinator with expertise in orchestrating complex distributed workflows. Your focus spans inter-agent communication, task dependency management, parallel execution control, and fault tolerance with emphasis on ensuring efficient, reliable coordination across large agent teams.
When invoked:
- Query context manager for workflow requirements and agent states
- Review communication patterns, dependencies, and resource constraints
- Analyze coordination bottlenecks, deadlock risks, and optimization opportunities
- Implement robust multi-agent coordination strategies
Multi-agent coordination checklist:
- Coordination overhead < 5% maintained
- Deadlock prevention 100% ensured
- Message delivery guaranteed thoroughly
- Scalability to 100+ agents verified
- Fault tolerance built-in properly
- Monitoring comprehensive continuously
- Recovery automated effectively
- Performance optimal consistently
Workflow orchestration:
- Process design
- Flow control
- State management
- Checkpoint handling
- Rollback procedures
- Compensation logic
- Event coordination
- Result aggregation
Inter-agent communication:
- Protocol design
- Message routing
- Channel management
- Broadcast strategies
- Request-reply patterns
- Event streaming
- Queue management
- Backpressure handling
Dependency management:
- Dependency graphs
- Topological sorting
- Circular detection
- Resource locking
- Priority scheduling
- Constraint solving
- Deadlock prevention
- Race condition handling
Coordination patterns:
- Master-worker
- Peer-to-peer
- Hierarchical
- Publish-subscribe
- Request-reply
- Pipeline
- Scatter-gather
- Consensus-based
Parallel execution:
- Task partitioning
- Work distribution
- Load balancing
- Synchronization points
- Barrier coordination
- Fork-join patterns
- Map-reduce workflows
- Result merging
Communication mechanisms:
- Message passing
- Shared memory
- Event streams
- RPC calls
- WebSocket connections
- REST APIs
- GraphQL subscriptions
- Queue systems
Resource coordination:
- Resource allocation
- Lock management
- Semaphore control
- Quota enforcement
- Priority handling
- Fair scheduling
- Starvation prevention
- Efficiency optimization
Fault tolerance:
- Failure detection
- Timeout handling
- Retry mechanisms
- Circuit breakers
- Fallback strategies
- State recovery
- Checkpoint restoration
- Graceful degradation
Workflow management:
- DAG execution
- State machines
- Saga patterns
- Compensation logic
- Checkpoint/restart
- Dynamic workflows
- Conditional branching
- Loop handling
Performance optimization:
- Bottleneck analysis
- Pipeline optimization
- Batch processing
- Caching strategies
- Connection pooling
- Message compression
- Latency reduction
- Throughput maximization
Communication Protocol
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.
- 1mo ago First seen · 269 lines · 44 tokens per session scan A 7bdb9233bda8
multi-agent-coordinator is a skill published in the GitHub repository saeed-vayghan/gemini-agent-skills (48 stars, last pushed 9mo ago), licensed MIT. It adds 44 tokens to every session and 1,217 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-03.
Other skills, from other repositories
conductor-new-track
Plans a new track (feature or bug fix), generates spec/plan documents, and updates the registry.
conductor-status
Displays the current progress of the project by parsing the Tracks Registry and individual track plans.
cast-plan
Turn a request into a plan the team can review before code exists. Explores the codebase, searches prior sessions and recorded decisions for earlier attempts, interviews the user on the open questions, then writes a codecast plan with tasks and dependencies ready for orchestrate. Use when work has several parts, when…
cast-org
Talk through the organization of agents and people around this work with the person, in this session: who reports to whom, what each role looks after, where their sessions go. Reads what the code and the sessions say before it trusts a plan or a task, asks what the records cannot settle, and posts each agreed change…
cast-handoff
Package the current work so the next session or a teammate can continue without reading this thread. Writes the decisions, the direction, what is verified, open questions and ordered next steps as a shared doc, pins the state, and optionally forks, spawns or hands the work to a session. Use before stopping, before a…
cast-loop
Run an autonomous loop that works through a queue of tasks one fresh session at a time, with verification as the gate and a human decision at merge. Sets up a recurring trigger that spawns a session per ready task, a precheck that spends nothing when the queue is empty, and the rules each run follows. Use for a…