devtu-optimize-skills

What problem does it solve?

This Skill provides a robust blueprint for optimizing ToolUniverse skills to produce high-quality, evidence-graded reports with thorough source attribution. It guides Skill authors through pre-flight tool verification, foundation data integration, disambiguation-first search, structured reporting, and explicit handling of data gaps and tool failures.

Core Features & Use Cases

• Tool Interface Verification: Before calling any tool, verify correct parameter usage and maintain a corrections table to prevent silent failures.
• Foundation Data Layer: Use a comprehensive aggregator (when available) first to establish a consistent baseline, then augment with specialized queries.
• Versioned Identifier Handling: Resolve both versioned and unversioned IDs and document which form works across APIs.
• Disambiguation Before Research: Resolve official IDs, gather synonyms, detect collisions, and collect baseline annotation data.
• Report-Only Output: Separate narrative findings from data; provide a dedicated bibliography JSON/CSV for all sources.
• Evidence Grading & Quantified Completeness: Apply T1-T4 evidence tiers and numeric minimums for each section to ensure objective completeness.
• Data Gaps & Retry/Fallback: Aggregate data gaps in a single section; implement retry logic and fallback chains for flaky tools.

Use Case Examples

• When creating a new ToolUniverse skill, run this optimization to produce a complete, evidence-graded metadata package.
• When auditing an existing Skill, apply the Phase -1 to Phase 3 workflow to improve structure and transparency.

Quick Start

Follow the Quick Start guidance to apply optimization on a target Skill unit, generating a metadata block compatible with the Skill Unit metadata schema.