What specifically does this skill analyze?

It analyzes any script execution using the 'python3 .plan/execute-script.py' pattern that returns a non-zero exit code, tracing the failure back to the specific instruction that triggered it.

How does the skill identify the root cause of a failure?

It compares the actual script call against the instructions found in the source component file to categorize the error, such as 'LLM Invented Script', 'Missing API', or 'Wrong Script Parameters'.

Does it work with any script or just Plan Marshall scripts?

It is optimized for the Plan Marshall framework's script notation and orchestration patterns, focusing on scripts called via the .plan directory.

Can I automatically apply fixes suggested by the skill?

Yes, the skill provides an interactive resolution step where you can choose to apply a proposed fix directly to the relevant component file using Claude's editing tools.

Script Failure Analyzer

Name: Script Failure Analyzer
Author: cuioss

bycuioss

0•

Security & Testing

Analyzes script failures to identify root causes and provide automated fix suggestions for AI orchestration workflows.

The Analyze Script Failures skill is a diagnostic powerhouse for the Plan Marshall orchestration layer, designed to handle non-zero exit codes during script execution. It automatically traces failed calls back to their source—whether a command, agent, or skill—and evaluates the instruction path to determine if the error was due to missing documentation, incorrect parameters, or LLM hallucinations. By providing detailed reports and interactive resolution options, it allows developers to quickly apply fixes to components or record insights into a 'Lessons Learned' database, ensuring continuous improvement in AI-driven development.

Key Features

01Automated detection of script execution errors from chat history

02Source tracing to pinpoint specific components and instruction lines

03Root cause categorization including missing APIs and parameter errors

04Interactive fix proposals for rapid component remediation

05Seamless integration with the Lessons Learned management system

060 GitHub stars

Use Cases

01Debugging why an AI agent incorrectly invoked a workspace script

02Standardizing the reporting of technical debt and script bugs during development

03Identifying missing documentation in marketplace bundles that leads to execution errors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cuioss/plan-marshall analyze-script-failures

For use in Claude.ai and ChatGPT

Download Skill