Skip to content
Get packs

Goal-Driven Finance Research

Tracks multi-step finance research against stated criteria and linked evidence, research only, no trading.

Source Download skill
Repository
HKUDS/Vibe-Trading
Stars
37,433
Licence
MIT
Files
1 file

What it does

When installed, the agent treats a complex finance research question as a tracked goal. It writes down a research-only objective and a set of acceptance criteria, then links each step (market data lookups, backtests, document reads, web sources) to the criterion it supports. It only marks the goal complete when every required criterion has supporting evidence, and otherwise reports it as blocked or lacking evidence. It is explicitly designed not to place or execute trades.

Great for

  • Analysts comparing strategies, assets, or market regimes who want an auditable trail of evidence
  • Researchers reviewing a investment thesis who need to see which criteria remain unresolved
  • Developers building finance research workflows on the Vibe-Trading platform

Skip it if

  • Anyone wanting live trade execution or order placement: the skill forbids it and refuses such instructions
  • Quick one-shot finance questions: the skill itself says not to open a goal for these, and the overhead is unnecessary
  • Investment decisions treated as advice: the skill sets a non-advice boundary, so outputs are research analysis, not recommendations

Try asking

  • “Compare the 2023 and 2024 performance of two momentum strategies and keep working until the conclusion is supported by backtest evidence”
  • “Audit my thesis that small-cap value outperforms in rising-rate regimes, with criteria for benchmarks and data dates”
  • “Keep working on this multi-asset comparison across turns and tell me which criteria are still unresolved”
Step-by-step guide: how to use Goal-Driven Finance ResearchInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    research-goal.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip research-goal.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip research-goal.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent concrete criteria up front (benchmark, symbol universe, date range) so the audit has something specific to check.
  • Ask for artifact-backed evidence such as saved backtest runs, since completion requires verified run IDs or hashed artifacts.
  • Expect the agent to report unresolved criteria rather than a confident answer when evidence is missing or stale.

What's inside

  • SKILL.md

Redistributed unmodified from HKUDS/Vibe-Trading under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.