trade-idea-generation
Systematic trade idea generation framework, fundamental quantitative and qualitative screening pipeline, catalyst-driven trade structuring, macro-to-micro idea translation, 20-60 day long-short portfolio management process
agentii-ai · 已有 AI 审核记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add agentii-ai/agentii-investment-intelligence --skill trade-idea-generation技能说明
name: trade-idea-generation description: Systematic trade idea generation framework, fundamental quantitative and qualitative screening pipeline, catalyst-driven trade structuring, macro-to-micro idea translation, 20-60 day long-short portfolio management process multi_ticker_semantics: single_target temporal_scope: default_quarters: 4 max_quarters: 12 description: "20-60 trading day horizon for trade idea generation; up to 12 quarters for macro context." allowed_tools:
- search_investment_strategies
- get_investment_strategy
- search_investment_cases
- get_investment_case
- search_by_analogue retrieval_scope: structured_only layer_tags: ["L2", "L3"] min_tool_diversity: 3 parameter_free: false
Methodology fused from professional trading and investment frameworks; all text is an original paraphrase.
Defaults
| Parameter | Default Value | Rationale |
|---|---|---|
| trade_horizon_days | 20-60 | Professional long/short portfolio management sweet spot |
| volatility_target | 20% annualized | Target portfolio volatility (1.5-2x VIX) |
| screening_universe | S&P 500 + liquid mid-caps | Default screening universe for trade ideas |
| portfolio_bias_source | macro-analysis skill | Derive bias from upstream macro view |
Preflight
Run canonical pre-flight per contracts/preflight.md. Propagate X-Agentii-Trace per contracts/x-agentii-trace-header.md.
Data Source Priority
- Process framework — embedded in this Protocol (proprietary systematic process)
- Quantitative data —
search_investment_strategies+ financial data from market data scripts - Qualitative context —
search_investment_cases+search_by_analoguefor historical analogues - Real-time context — market conditions and sector trends from data tools
Methodology
Retrieval Scope
structured_only
Retrieval Strategy
Ownership & insider signals: search_institutional_holdings (top-10 holders + whale portfolios, direction=accumulating|reducing|new|exited) and search_insider_trades (Form-4 transactions with SEC URLs) are available as signal inputs.
This skill follows Branch (d) Simple Lookup from contracts/retrieval.md: query search_investment_strategies for fundamental analysis frameworks, search_investment_cases for historical case analogues, search_by_analogue for cross-strategy discovery. Delegates detailed analysis to sub-skills. No unstructured document retrieval.
Temporal Scope
See frontmatter temporal_scope block.
Tool Allowlist
See frontmatter allowed_tools.
Protocol
This skill orchestrates a systematic trade idea generation pipeline for the 20-60 day long/short portfolio management horizon. Detailed methodology, funnel architecture, composite scoring, operational cadence, and the trade-vs-investment hard gate are in references/pipeline-orchestration.md.
Core formula: Trade Idea = Fundamentals + Timing + Trade Structure. If any component is missing, the idea is not actionable.
Hard gate: A trade idea requires a specific, dateable catalyst within 20-60 days. Without one, the idea is an investment, not a trade — move to watchlist.
Pipeline funnel (6 gates, see reference for thresholds):
Universe → Gate 1: Quant Screen → Gate 2: Financial Validation →
Gate 3: Qualitative Filter → Gate 4: Template Assembly →
Gate 5: Strategy Match → Gate 6: Trade Structure → Pipeline Ranking
Composite: (Quant × 0.30) + (Qual × 0.30) + (Catalyst × 0.25) + (Macro × 0.15), each 1-10.
Phase 1 — Theory
- Macro Bias Input: Receive portfolio bias (long/short/neutral) from upstream. Long → $3B-$10B mid-caps. Short → $20B+ large caps (never short mid-caps — takeover risk). Neutral → emphasize pair trades. Assess sector tailwinds/headwinds.
- Quantitative Screening: Delegate to
quantitative-screening. Inputs: bias, sectors, universe. Outputs: ranked outlier list, financial validation, growth profiles, trap classification, data quality flags. - Qualitative Deep Dive: Delegate to
qualitative-filteringfor top <10 candidates. Inputs: candidate list, metrics, sector context. Outputs: KPI analysis, MOP credibility (5-dimension scorecard), management/board assessment, catalyst calendar with specificity/magnitude/probability. - Template Assembly: Delegate to
trade-template. Inputs: all quant/qual outputs, macro, peer data. Outputs: multi-method price targets, probability-weighted scenarios (Base 55/Bull 20/Bear 25), GAAP reconciliation, thesis statement. - Macro-Driven Cases: Delegate to
macro-idea-generationfor top-down ideas. Inputs: regime, indicators, international context. Outputs: regime-to-sector mapping, macro catalyst timeline, ADR candidates, invalidation thresholds.
Phase 2 — Implementation
- Strategy Matching:
search_investment_strategies(domain=fundamental)→search_by_analogue(company_situation, event_type)→get_investment_strategy. Cross-reference: do matched strategies confirm or challenge the thesis? - Historical Precedent:
search_investment_cases(domain=[derived]). Flag parallels and divergences. Cite via/v/links. Note gaps explicitly — absent precedent is itself information. - Trade Structure: Stock (default), options (delegate to
options-execution), or pair trade (market-neutral, macro divergence). Requirements: identifiable catalyst within 60 days, R/R > 2:1. - Pipeline Output: Rank by composite. Classify: Actionable (all gates, catalyst < 60d, R/R > 2:1) / Watchlist (catalyst > 60d, R/R borderline) / Discard (failed gates 1-3). Document full audit trail.
Output File
{ticker}/{YYYY-MM-DD_HHMM}_trade-idea-generation_{affix}.md
Output Structure
- Executive Summary — Pipeline funnel statistics (universe → gate 1 → ... → gate 6 counts), composite scores, top 3 actionable recommendations
- Macro Context — Portfolio bias (long/short/neutral), macro regime, sector tailwinds/headwinds
- **Quantitative Scree