factor

Orchestrates factor research by composing DataAccess, FactorEngine, and QuantEvaluator capabilities.

Updated Jun 27, 2026
One-click install
npx skills add https://github.com/HKUST-QUANT-SOCIETY/quantcode --skill factor-hkust-quant-society
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: factor
Source: https://github.com/HKUST-QUANT-SOCIETY/quantcode/tree/main/.opencode/groups/factor/skills/factor
Command: npx skills add https://github.com/HKUST-QUANT-SOCIETY/quantcode --skill factor-hkust-quant-society

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Quantitative factor research often fragments across disconnected tools for data access, factor computation, and evaluation, leading to duplicated implementations and unverified results. This Skill acts as the factor group's orchestration agent, discovering and reusing the organization's canonical DataAccess, FactorEngine, QuantEvaluator, and FactorAssets components instead of rewriting them. ## Core Features & Use Cases - Capability Discovery and Reuse: Loads session context, queries the Capability Catalog, group Memory, and data contracts before acting, ensuring work builds on published components. - Contract-Verified Factor Proposals: Drafts a FactorSpec, then validates it through published contract-verification capabilities so model-generated output is never treated as verified. - Governed Evaluation and Sharing: Routes evaluation through the canonical QuantEvaluator, checks ComponentCallResult environment, version, and provenance, and only shares results through the approved merge workflow with receipts. - Use Case: A factor group researcher asks the agent to evaluate a new momentum factor; the agent locates the right DataAccess and FactorEngine coverage, runs the canonical QuantEvaluator, and proposes an L2/L3 storage plan awaiting task-level approval. ## Quick Start Ask the factor agent to draft a FactorSpec for your factor idea and evaluate it with the canonical QuantEvaluator using authorized workspace data.

Frequently Asked Questions about factor

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

FAQPage Schema
How do I evaluate a new factor with the factor group agent?

Draft a FactorSpec with the agent, which then validates it through published contract-verification capabilities and runs the canonical QuantEvaluator. Model-generated output is never treated as verified until contract verification and evaluation results confirm it.

What components does the factor skill reuse instead of rewriting?

It reuses the organization's canonical DataAccess, FactorEngine, QuantEvaluator, FactorOptimizer, and FactorAssets components. Rewriting these inside QuantCode is explicitly forbidden; the agent selects components based on coverage of the task.

Can the factor agent share results to shared storage automatically?

No. Sharing requires the approved merge workflow with exact merge approval and receipts. L2/L3 writes must be proposed through organization_solution and wait for the task interface's freeze decision on the current version.

What happens when a required component is unavailable?

The agent returns UNAVAILABLE when a component is not connected. Proxy, mock, or staging components may only be used for development regression, never as production evidence or to trigger automatic ingestion.

What does the factor agent do when capabilities only partially cover a task?

It reports the coverage gap to the user and asks whether to adapt, clarify, or build a custom implementation. It does not silently rebuild missing functionality on its own.