How does the skill identify cost outliers?

It compares actual task costs against pre-defined benchmarks for different model complexities (Haiku, Sonnet, and Opus) to flag unexpected overruns.

Can this skill fix failed tasks automatically?

It identifies issues and proposes specific follow-up actions, such as retrying with higher complexity or splitting tasks, which the user can then approve to execute.

What is the difference between Full Review and Quick Status?

Full Review performs a deep 6-phase analysis including pattern detection and report generation, while Quick Status provides a fast, read-only snapshot of progress and costs.

Does it review the actual code changes?

Yes, during Full Review mode, the skill can optionally perform a git diff analysis of the changes made during the run to check for issues or missing tests.

What files does this skill require to function?

It primarily analyzes the .maximus/progress.md and .maximus/run-summary.json files generated by the Maximus Loop engine.

Maximus Review Analyzer

Name: Maximus Review Analyzer
Author: itsdevcoffee

byitsdevcoffee

•

Analytics & Monitoring

Analyzes Maximus Loop execution results to provide performance insights, cost breakdowns, and actionable follow-up recommendations.

Maximus Review is a specialized diagnostic skill for Claude Code designed to evaluate the outcomes of Maximus Loop runs. It provides deep visibility into task execution by parsing progress metadata and run summaries to identify failure patterns, cost outliers, and complexity mismatches. Whether performing a comprehensive six-phase analysis or a quick status check, this skill helps developers optimize their AI-driven workflows by suggesting retry strategies, task splitting, and plan extensions based on empirical execution data.

Key Features

01Automated six-phase execution analysis with visual progress tracking

021 GitHub stars

03Interactive code change reviews via integrated git diff analysis

04Detailed cost-per-task breakdown against Haiku, Sonnet, and Opus benchmarks

05Intelligent pattern detection for timeouts, retries, and complexity mismatches

06Actionable follow-up proposals for task retries and plan extensions

Use Cases

01Evaluating the success rate and total cost of a long-running autonomous coding task

02Diagnosing why specific tasks in a multi-step plan are failing or timing out

03Monitoring the progress and estimated completion time of an active execution loop

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add itsdevcoffee/devcoffee-agent-skills maximus-review

For use in Claude.ai and ChatGPT

Download Skill