Record observable test runs and write a Markdown debug report.
npx skills add https://github.com/shenmuxing/EtaSkill --skill skill-tester-shenmuxing
Please help me install this Agent Skill. Skill: skill-tester Source: https://github.com/shenmuxing/EtaSkill/tree/main/skill-examples/skill-tester Command: npx skills add https://github.com/shenmuxing/EtaSkill --skill skill-tester-shenmuxing
This Skill helps you inspect how another Codex skill behaves on a real request by producing a local debug report of the run, including visible actions, files read, commands executed, and any errors or artifacts.
Use the skill-tester skill to test the target skill on my request and write a local debug report.
High-intent search queries and answers about installing and using this skill.
To debug a Codex skill run, you can use a skill tester to capture observable actions like files read and commands run, then write a local Markdown debug report documenting the behavior for troubleshooting.
A dry run for Codex skills is a behavior inspection process that captures steps taken during a concrete skill run. You need it when a skill seems unclear or produces unexpected results and you want a reproducible record.
You generate a Markdown debug report by running a skill tester against a target skill on a specific request, which records inputs, outputs, commands executed, changed artifacts, and errors into a non-overwriting local file.
Yes, you can compare a Codex skill's behavior against a concrete user request by running an observable test that documents the skill's decisions, visible actions, and changed artifacts in a local debug report.
No, the skill testing report does not overwrite previous runs. It writes a non-overwriting Markdown file to ensure you retain a reproducible record of every debug session and behavior inspection.
If a Codex skill behaves unexpectedly, run an observable skill test on your request to capture visible actions, errors, and changed artifacts, then review the generated Markdown debug report to understand the behavior.
yyoonchulTransform user prompts into ready-to-run Agent Skill scaffolds.
lee-toOrchestrate iterative artifact refinement through PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE phases.
hajrasarwar11Create and manage persistent project tasks with .local/tasks markdown plan files.
kk20300113-pngDefine pass/fail criteria and calculate pass@k metrics for Claude Code agent evals.
