add-problem-instance

Adds executable optimization problem instances as paired JSON metadata and Python registry entries.

Updated Jul 13, 2026
One-click install
npx skills add https://github.com/mryk814/optimization-compass --skill add-problem-instance-mryk814
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: add-problem-instance
Source: https://github.com/mryk814/optimization-compass/tree/main/.agents/skills/add-problem-instance
Command: npx skills add https://github.com/mryk814/optimization-compass --skill add-problem-instance-mryk814

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Adding a new benchmark problem instance to the Optimization Compass knowledge base requires coordinated edits across two authoritative files, and changing only one leaves the dataset inconsistent and failing validation. ## Core Features & Use Cases - Paired Authority Updates: Guides synchronized edits to problem-suite.json (metadata) and problem_registry.py (evaluator), whose registry keys must match exactly. - Validation Workflow: Enforces running uv run optimization-compass validate problem and the Tier C PR gate before submission. - Test Pattern Guidance: Directs adding focused tests following existing patterns in tests/test_problem_instances.py. - Use Case: When contributing a new benchmark instance (e.g., a knapsack variant) to the optimization knowledge base, follow this Skill to mirror the closest existing mathematical family, add tests, and pass the Tier C validation gate. ## Quick Start Add a new executable problem instance to the optimization-compass problem suite following the paired JSON and registry workflow.

Frequently Asked Questions about add-problem-instance

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

FAQPage Schema
How do I add a new problem instance to optimization-compass?▼

Add paired entries to problem-suite.json for metadata and problem_registry.py for the evaluator, keeping registry keys exactly matched. Then add a focused test in tests/test_problem_instances.py and run uv run optimization-compass validate problem.

What validation is required before submitting a problem instance PR?▼

The PR gate is uv run optimization-compass validate tier-c, since executable problems fall under Tier C. Focused validate problem or tier-b runs do not replace this gate, and DCO sign-off is mandatory.

Why must problem-suite.json and problem_registry.py be changed together?▼

The instance registry_key in the JSON metadata must exactly match the key set in the Python registry. Editing only one authority leaves the pair inconsistent and fails validation.

When should I stop instead of adding a problem instance?▼

Stop if there is no reference source grounding the instance, following the maintenance skill's stop conditions and the knowledge-change checklist. Also decide first whether an existing definition suffices before creating a new one.