workflow-optimizer

Analyzes agent tool-calling patterns and task execution efficiency to suggest workflow improvements.

Updated Apr 13, 2026
One-click install
npx skills add https://github.com/JenilRevaliya/ARGUS --skill workflow-optimizer-jenilrevaliya
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: workflow-optimizer
Source: https://github.com/JenilRevaliya/ARGUS/tree/main/.agent/skills/workflow-optimizer
Command: npx skills add https://github.com/JenilRevaliya/ARGUS --skill workflow-optimizer-jenilrevaliya

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents often waste tool calls through redundant file reads, serial execution of independent operations, and inefficient error recovery, inflating task time and context usage. This Skill audits actual tool call logs to identify waste and recommend concrete optimizations. ## Core Features & Use Cases - Tool Call Pattern Analysis: Classifies calls into waste patterns like redundant reads, blind searches, serial bottlenecks, and ping-pong edits with severity ratings. - Parallelism Detection: Identifies independent tool calls that should run concurrently based on data dependency rules. - Efficiency Scoring: Computes an A-F graded efficiency score from actual versus optimal tool call counts with a parallelism penalty adjustment. - Use Case: After a complex multi-step orchestration task, review the execution trace to find that a file was read three times and four independent reads ran serially, then receive a report showing how to cut 14 calls down to 9. ## Quick Start Analyze the tool call history from my last task and generate a workflow optimization report with efficiency grade and recommendations.

Frequently Asked Questions about workflow-optimizer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze AI agent tool call efficiency?▼

Provide the actual sequence of tool calls from a completed task and the Skill classifies each call into waste patterns like redundant reads or serial bottlenecks. It then computes an efficiency score comparing actual calls to the estimated optimal count.

How to reduce the number of tool calls an agent makes?▼

Batch independent file reads into parallel calls, use multi-edit operations instead of sequential edits on the same file, and read only targeted code sections instead of full files. The Skill's quick win checklist covers the most common savings.

What is a good efficiency score for an agent workflow?▼

Scores of 90-100% earn grade A and indicate near-optimal execution, while 75-89% is grade B with minor opportunities. Below 60% signals significant waste requiring workflow strategy rework.

Can this analyze tool calls that never happened?▼

No. The anti-hallucination guard requires analysis of actual tool call logs only, and recommendations must reference tools available in the current environment. Efficiency scores are always calculated from real call counts.

When should workflow optimization analysis be run?▼

Run it when tasks take more tool calls than expected, after complex orchestration or swarm dispatches, during retrospectives, or when context window pressure causes truncated responses and missed context.