agent-orchestration-multi-agent-optimize

Optimize multi-agent systems with coordinated profiling, workload distribution, and cost-aware orchestration. Use when improving agent performance, throughput, or reliability.

Install with Tessl CLI

npx tessl i github:sickn33/antigravity-awesome-skills --skill agent-orchestration-multi-agent-optimize

What are skills?

3.00x

Quality

47%

Does it follow best practices?

Impact

96%

3.00x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/agent-orchestration-multi-agent-optimize/SKILL.md

SKILL.md

Review

Evals

Evaluation results

100%

52%

ShopFast Performance Profiling Framework

Multi-layer profiling agents

Criteria

Without context

With context

DatabasePerformanceAgent class

100%

ApplicationPerformanceAgent class

100%

FrontendPerformanceAgent class

100%

All three agents instantiated

100%

aggregate_performance_metrics call

100%

Per-agent profile collection

50%

100%

Database-layer metrics

100%

Application-layer metrics

100%

Frontend-layer metrics

100%

Baseline establishment

100%

Validation with rollback

100%

Without context: $0.6109 · 2m 51s · 19 turns · 18 in / 11,657 out tokens

With context: $0.8104 · 3m 33s · 27 turns · 74 in / 12,977 out tokens

100%

58%

MLOps Co: LLM Budget Management System

Cost-aware model selection

Criteria

Without context

With context

CostOptimizer class

100%

token_budget value

100%

gpt-5 model cost

100%

claude-4-sonnet model cost

100%

claude-4-haiku model cost

100%

select_optimal_model method

100%

Budget-aware model selection

100%

token_usage tracking

100%

Incremental rollout plan

100%

Rollback mechanism

100%

Without context: $0.3574 · 1m 55s · 12 turns · 61 in / 6,714 out tokens

With context: $0.5797 · 2m 53s · 21 turns · 69 in / 9,067 out tokens

90%

82%

AgentFlow: Parallel Orchestration Pipeline Redesign

Parallel orchestration with context compression

Criteria

Without context

With context

MultiAgentOrchestrator class

100%

PriorityQueue for execution_queue

100%

PerformanceTracker instance

100%

ThreadPoolExecutor usage

100%

as_completed pattern

100%

performance_tracker.log call

100%

semantic_truncate function

100%

max_tokens=4000 parameter

100%

importance_threshold=0.7 parameter

100%

Caching or memoization

Fault tolerance handling

100%

Without context: $0.4859 · 2m 34s · 15 turns · 16 in / 9,000 out tokens

With context: $0.7501 · 3m 10s · 28 turns · 75 in / 10,455 out tokens

Evaluated: 8 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

ShopFast Performance Profiling Framework MLOps Co: LLM Budget Management System AgentFlow: Parallel Orchestration Pipeline Redesign

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.