# Optimize Multi-Agent System Performance

Multi-agent workflows often waste time, tokens, and compute through poor coordination. This skill guides profiling, workload distribution, cost controls, and measured optimization.

## Install

```bash
npx skillstore add sickn33/agent-orchestration-multi-agent-optimize
```

## Metadata

- Status: approved
- Slug: sickn33-agent-orchestration-multi-agent-optimize
- Skillstore revision: r2
- Version status: missing
- Tree hash: d4659c7c3277e8917452462f67c36ec61829abcd3b8bf1ff03b6eaf12814e496
- Author: sickn33
- GitHub username: sickn33
- License: MIT
- Repository: https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/agent-orchestration-multi-agent-optimize
- Ref: ebdfe608f5de2b66ff37ab4af12af8ac4f5e8006
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands
- Quality score: 79
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/sickn33-agent-orchestration-multi-agent-optimize
- Manifest: https://skillstore.pages.dev/api/skills/sickn33-agent-orchestration-multi-agent-optimize/manifest

## Capabilities

- Defines baseline metrics and target performance goals before optimization.
- Profiles database, application, and frontend agent workloads for bottlenecks.
- Recommends context compression, relevance filtering, and token budget controls.
- Designs parallel execution and dynamic workload distribution strategies.
- Tracks token usage and supports cost-aware model selection.
- Guides gradual rollouts with repeatable tests and rollback planning.

## Use Cases

- Reduce Workflow Latency: Profile agent handoffs, identify blocking work, and design parallel execution with measurable latency targets.
- Control LLM Costs: Track token use, select models by task complexity, and define budget-aware context policies.
- Improve Agent Reliability: Plan incremental orchestration changes, regression tests, quality thresholds, and rollback criteria.

## Prompt Templates

### Establish a Baseline

```
Review this multi-agent workflow: [workflow]. Identify measurable latency, throughput, quality, and cost baselines before proposing changes.
```

### Profile Coordination Bottlenecks

```
Profile [system] against [performance goals]. Map each agent's workload, dependencies, communication overhead, context use, and likely bottlenecks.
```

### Design an Optimization Plan

```
Design an incremental orchestration plan for [system] under [budget constraints]. Include workload distribution, parallelism, context controls, rollback steps, and evaluation metrics.
```

### Evaluate Competing Strategies

```
Compare two orchestration strategies for [system] using [evaluation data]. Quantify cost, latency, throughput, reliability, and quality tradeoffs, then recommend a staged rollout.
```

## Limitations

- Provides guidance and illustrative pseudocode, not a complete executable framework.
- Requires baseline metrics and evaluation data for meaningful recommendations.
- Does not focus on tuning a single agent prompt.
- Requires environment-specific testing before production deployment.

## Best Practices

- Measure performance, quality, and cost before and after every optimization.
- Change one orchestration variable at a time when practical.
- Use staged rollouts, explicit thresholds, and tested rollback procedures.

## Anti Patterns

- Do not optimize without baseline metrics or representative evaluation data.
- Do not parallelize tasks that have unresolved dependencies or shared-state conflicts.
- Do not reduce cost by accepting unmeasured quality degradation.

## Security Audit

- Audited at: 2026-07-23T20:26:22.282\+00:00
- Summary: All 13 static findings are false positives caused by Markdown placeholders, code fences, and a deployment safety warning. No executable shell commands, reconnaissance behavior, prompt injection, or malicious intent appears in SKILL.md.

## Stats

- Views: 158
- Downloads: 12
- Favorites: 0
- Popularity score: 0