Verified:GitHub

skill-optimizer

Diagnose Agent Skill failures and generate reviewable patches from execution evidence.

280|30|Updated Mar 30, 2026

Search Tags:#benchmarking #agent skills #packaging #patch generation #trigger evaluation #evidence-based optimization #validation cases

Ready to run

npx skills add https://github.com/Undertone0809/rudder --skill skill-optimizer-undertone0809

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-optimizer
Source: https://github.com/Undertone0809/rudder/tree/main/server/resources/bundled-skills/skill-optimizer
Command: npx skills add https://github.com/Undertone0809/rudder --skill skill-optimizer-undertone0809

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Skill Optimizer reduces unreliable or wrongly-triggered Agent Skill behavior by converting real-world conversation and execution evidence into safer, clearer, reviewable updates.

Core Features & Use Cases

Evidence-based diagnosis: identifies what observations imply about root causes and what must change to prevent repeats.
Reviewable patch generation: proposes small, auditable edits to a target skill, including updated inputs, triggers, workflow steps, and outputs.
Validation and benchmarking: produces validation cases and trigger eval suggestions to verify improvements and catch regressions.
Packaging support: organizes skill folders and associated references, scripts, and documentation into a distributable structure.

Quick Start

Use the skill-optimizer when you want to improve an existing Agent Skill (for example, “Optimize the release-skill SKILL.md based on a failed eval trace and user corrections.”).

Frequently Asked Questions about skill-optimizer

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

FAQPage Schema

How do I fix an Agent Skill that triggers incorrectly or behaves unreliably?▼

Fix unreliable Agent Skill behavior by applying evidence-based diagnosis to real-world conversation and execution traces, identifying root causes, and generating reviewable patches that update triggers, inputs, and workflow steps to prevent repeat failures.

What is evidence-based skill hardening and when do I need it?▼

Evidence-based skill hardening converts observed failures and corrections into safer, clearer, reviewable updates. You need it when an existing Agent Skill requires debugging, generalization, or durable changes derived from execution evidence.

How do I package an Agent Skill for distribution?▼

Package an Agent Skill by organizing skill folders, associated references, scripts, and documentation into a distributable structure, ensuring the hardened skill and its validation cases are ready for deployment.

How do I benchmark and validate an Agent Skill to catch regressions?▼

Benchmark and validate an Agent Skill by producing validation cases and trigger evaluation suggestions, verifying that evidence-based patches improve behavior and catch regressions in the updated skill workflow.

Can I generate reviewable patches for SKILL.md without breaking safety constraints?▼

Yes, you can generate reviewable patches for SKILL.md sections while preserving safety constraints and appropriate authority gates. Proposed edits are small, auditable, and include updated inputs, triggers, workflow steps, and outputs.