fundamental-compose

Orchestrates fundamental research tasks using authorized PIT data, financial components, and report rendering.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Conducting fundamental equity research requires combining point-in-time data, financial extraction, valuation models, and report generation while respecting organizational permissions and data authorization rules. This Skill coordinates that entire workflow under the fundamental group's identity and governance constraints. ## Core Features & Use Cases - PIT-compliant research: Enforces published_at <= as_of_date on all source materials, preserving citations with source and timestamp. - Component-based analysis: Invokes published capability adapters such as pit_rag_search, extract_financial, dcf_valuation, and render_report when available in the service catalog. - Governed report production: Uses organization_solution to define file scope and acceptance criteria before writing or rendering reports, with honest reporting of any unavailable or failed components. - Use Case: A fundamental group analyst needs a valuation report on a company as of a specific date. The Skill queries the capability catalog and group Memory, reads authorized PIT materials, runs financial extraction and DCF valuation components, then renders a report for user acceptance. ## Quick Start Start a fundamental research task for a target company with a specified as_of_date, asking it to query the capability catalog and group Memory first, then produce a cited valuation report using only authorized PIT sources.

Frequently Asked Questions about fundamental-compose

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

FAQPage Schema
How do I run a fundamental research task with PIT data compliance?

Confirm the target company or industry and the as_of_date first, then read only authorized materials where published_at is on or before that date. Keep source and timestamp citations for every reference used in the final report.

What components does a fundamental research workflow use?

The capability catalog may publish adapters such as pit_rag_search, extract_financial, dcf_valuation, and render_report. Actual availability depends on the server-published catalog and permissions, so adapter names alone do not guarantee connectivity.

Can I use fixture or stub data for valuation results?

No. Financial data and valuation calculations must come from real authoritative components. Fixtures, stubs, or hash-generated numbers cannot be presented as research conclusions, and failed steps must be reported honestly.

How are research reports approved and published?

Report quality review happens inside the task, and file writes require an organization_solution defining scope and acceptance before execution. External submission needs an actually published sharing interface with explicit approval; producing a local PDF does not mean the report is published.

Does the research report require a separate model API key?

No. Report content is generated by the current task's host model based on the gathered materials. No separate Python model key is configured for report writing.