Bigdata.com

View saved version →

---
name: bigdata-catalyst-monitor
description: >
  Map the dated events that could move a public company over the next few quarters, using
  Bigdata.com data (events calendar, filings, news, tearsheet). Covers scheduled catalysts —
  earnings, investor days, index reviews, lock-up and patent expiries, regulatory decision
  dates — and foreseeable unscheduled ones — litigation milestones, product cycles, contract
  renewals, refinancings. Each catalyst carries a date or window, likely direction, magnitude,
  confidence, and what to watch, ranked by expected impact rather than by date alone.
  Triggers: "catalyst monitor for X", "what's coming up for X", "upcoming catalysts for X",
  "what could move X", "key dates for X", "event calendar for X", "what should I watch on X".
---

# Bigdata Catalyst Monitor

Forward calendar of what could move the name, ranked by impact. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the question is "what's coming". Not this skill when:

| Request | Use instead |
|---------|-------------|
| What already happened over the past month | Company brief |
| Deep analysis of the next earnings print | Earnings preview |
| Probability-weighted outcomes and values | Scenario analysis |
| Risks rated by likelihood and impact | Risk assessment |

**Catalyst vs risk:** a catalyst is a **dated or datable event** that resolves something. A standing risk with no resolution date belongs in a risk assessment.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_events_calendar` | Scheduled earnings and conferences | `find_securities` |
| `bigdata_company_tearsheet` | Baseline financials, estimates, what the price embeds | `find_securities` |
| `bigdata_search` | Regulatory dates, litigation milestones, product cycles, filings | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Scheduled events

Call `bigdata_events_calendar` for earnings dates and conferences over the next 2–4 quarters. Note the fiscal calendar so quarter-ends and guidance updates land correctly.

### Step 3 — Baseline: what is already priced

Call `bigdata_company_tearsheet`. A catalyst only matters relative to expectations — capture consensus estimates, current multiples, and sentiment so each catalyst can be framed against what the market already assumes.

### Step 4 — Search for datable events

Run **6–8 targeted searches**:

- "[Company] upcoming product launch roadmap timeline"
- "[Company] regulatory decision approval date FDA / FTC / EU"
- "[Company] lawsuit trial date court ruling expected"
- "[Company] investor day capital markets day guidance update"
- "[Company] contract renewal expiry major customer"
- "[Company] debt maturity refinancing schedule"
- "[Company] patent expiry exclusivity loss"
- "[Company] index inclusion review lock-up expiry"

Include the industry-specific ones that apply — clinical readouts, license renewals, rate cases, spectrum auctions, model launches.

### Step 5 — Rate each catalyst

For every catalyst record:

- **Date or window** — exact where known, quarter where not; mark undated but expected items as such
- **Type** — scheduled or foreseeable
- **Likely direction** — positive / negative / two-sided
- **Magnitude** — high / medium / low, tied to a value driver where possible ("~$Nm revenue", "~Xbps margin", "removes overhang on segment Y")
- **Confidence** — how sure the date and the outcome are; these are different, and both matter
- **What to watch** — the specific signal that tells you which way it resolved

### Step 6 — Rank and sequence

Rank by **expected impact**, not chronology — a large two-sided event in nine months usually matters more than a routine print next week. Then give the chronological calendar separately, so the reader gets both views.

Close with the **2–3 catalysts that dominate** the next few quarters and what each would change.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version if useful.

## Quality bar

Non-negotiables:

- Every catalyst carries a **date or window** — undated speculation is not a catalyst
- Direction, magnitude, and confidence given separately; date confidence distinguished from outcome confidence
- Magnitude tied to a value driver, not just labelled "high"
- Ranked by impact **and** listed chronologically
- Standing risks with no resolution date excluded — they belong in a risk assessment
- Facts separated from analysis and implications

Referenced files: 5

View saved version →

---
name: bigdata-company-brief
description: >
  Generate a company brief — a cited 30-day summary of what happened at a public company and
  why it matters — using Bigdata.com data (news, filings, transcripts, tearsheet financials).
  Findings are categorized into financial results, product and tech launches, M&A and
  partnerships, regulatory and legal, management changes, and other material events, each with
  date, facts, and a bullish/bearish/neutral investment implication tied to a value driver, plus
  competitive context and a ranked top 2-3. Triggers: "company brief for X", "what's happening
  with X", "catch me up on X", "recent developments at X", "what's the news on X",
  "summary of the last month for X", "any updates on X".
---

# Bigdata Company Brief

Retrospective 30-day summary of material developments at a public company, with a "so what" on each. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants to catch up on what has already happened. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Analysis ahead of an upcoming earnings call | Earnings preview |
| Breakdown of results just reported | Earnings digest / earnings reaction |
| "What is it worth" | Valuation snapshot |
| Structured risk mapping with likelihood/impact | Risk assessment |
| Full thesis, DCF, or variant perception | Investment memo |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Profile, sector/industry, financial position, recent performance | `find_securities` |
| `bigdata_search` | News, filings, transcripts, legal and regulatory coverage | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Business context

Call `bigdata_company_tearsheet` with the `entity_id` for the company profile (sector, industry, description), financial position, and recent performance metrics. This context is what lets you judge whether a news item is material — read it before searching.

### Step 3 — Competitive context (one pass)

Run `bigdata_search` at least once on industry structure and positioning:

- "[Company] competitive landscape market share"
- "[Company] vs competitors [Industry]"

A full five-forces write-up is not needed. **2–4 sentences** on concentration, pricing power, and disruption risk materially lift the brief. Mental model: [references/porter-five-forces.md](./references/porter-five-forces.md).

### Step 4 — Search recent news (last 30 days)

Use `bigdata_search` with natural-language queries that include the company name and a temporal reference. Run **5–10 searches** for coverage, then compress.

- "[Company] news last 30 days"
- "[Company] recent developments"
- "[Company] earnings announcement"
- "[Company] product launches partnerships"
- "[Company] regulatory legal updates"
- "[Company] lawsuit litigation court ruling investigation settlement last 30 days"

Cover all of: financial developments, product/technology announcements, partnerships and M&A, regulatory and legal matters, management changes. The explicit legal/litigation query is there so material non-operational items don't get buried under earnings headlines.

If the month was quiet, ask:

> "I haven't found any significant developments for [Company] in the last 30 days. Would you like me to extend the search period or focus on specific topics?"

### Step 5 — Categorize, and mark what is primary

Sort every finding into the six output categories: **Financial Results**, **Product/Tech Launches**, **M&A and Partnerships**, **Regulatory/Legal Updates**, **Management Changes**, **Other Material Events**.

While sorting, tag each item **primary** (material to value or the narrative) or **secondary**. Do not equal-weight the categories in the prose — quantity of headlines is not materiality. If a category is empty, write "No significant developments in this period" rather than padding it.

### Step 6 — Investment implication ("so what")

For each **material** event, give:

- **Date** — when it happened or was announced
- **Facts** — objective summary of what happened
- **Investment implication** — Bullish / Bearish / Neutral, tied to a value driver

Avoid generic labels. Replace "bullish for the stock" with the lever: "could support ~+$Nm revenue run-rate", "+Ybps margin", "de-risks [issue]", "lifts regulatory overhang on segment Z". If it isn't quantifiable, state the **specific mechanism** instead of a direction.

### Step 7 — Rank and synthesize

Rank the top **2–3** developments for the period and lead the executive summary with them. The overall assessment states the **net tilt and why**, not a recap of every bullet.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, sources, and footer.

- Add inline citations as superscript-style numbers `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version at the end if useful.

## Quality bar

Pass the PM test before delivering: **What's different?** **What matters (2–3 items)?** **What should I do about it?** (net assessment, key risk, next catalyst — no position sizing). Would it survive a short, skeptical morning meeting without reading as a news dump?

Non-negotiables in every brief:

- Competitive context present, not skipped
- Every material event carries date, facts, and a **mechanism-level** implication
- Top 2–3 developments ranked and surfaced in the executive summary
- Legal/regulatory search actually run, with material items shown
- Empty categories marked as such rather than padded
- Facts separated from analysis and implications

Referenced files: 5

View saved version →

---
name: bigdata-country-analysis
description: >
  Produce a deep country economic analysis using Bigdata.com data — GDP, inflation, monetary
  policy, labor markets, debt mechanics, and investment implications. Goes beyond a
  point-in-time snapshot: structural and historical context (sector transformation, labor
  productivity), debt composition and servicing, tax-to-GDP and public financial management, a
  substantive macro-and-micro labor section, market implications across equities, rates,
  currency and FDI, and a dedicated sourced policy-recommendations section. Suits institutional,
  multilateral, and academic audiences. Triggers: "economic outlook for X", "analyze X's
  economy", "country analysis of X", "how is X's economy doing", "X GDP inflation outlook",
  "X fiscal position", "X monetary policy outlook".
---

# Bigdata Country Analysis

Analytical country economic profile with policy depth. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the subject is one country's economy. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Several regions or blocs compared | Regional comparison |
| The G7 specifically | G7 comparison |
| A sector inside a country | Country-sector analysis |
| A cross-border macro theme | Thematic research |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_country_tearsheet` | Economic data, calendar, comparisons | None |
| `bigdata_search` | Indicators, policy, structural context, market implications | None |

If `bigdata_country_tearsheet` is unavailable or fails, complete the whole analysis with `bigdata_search` using the targeted queries below. Search **each indicator separately** rather than in one broad query.

## Workflow

### Step 1 — Core economic indicators

- "[Country] GDP growth economic outlook 2026"
- "[Country] inflation CPI consumer prices trends"
- "[Country] central bank interest rates monetary policy"
- "[Country] unemployment rate labor market"
- "[Country] fiscal policy government budget"

### Step 2 — Monetary policy

- "[Country] central bank rate decision outlook"
- "[Country] monetary policy inflation target"
- "Fed / ECB / BOJ / PBOC policy rate path expectations"

### Step 3 — Economic calendar and events

- "[Country] economic data releases calendar"
- "[Country] central bank meeting schedule"
- "[Country] GDP CPI employment report dates"

### Step 4 — Structural and historical context (required for depth)

A point-in-time-only report fails this deliverable's bar.

- "[Country] sector transformation structural change agriculture industry services"
- "[Country] labor productivity by sector comparison"
- "[Country] economic structure evolution [time range, e.g. 1990 2020]"
- "[Country] sectoral GDP share history"
- "[Country] employment by sector productivity growth"

### Step 5 — Debt composition, tax-to-GDP, and PFM (required for depth)

- "[Country] public debt composition domestic external"
- "[Country] debt servicing interest cost weighted average rate"
- "[Country] debt crowding out private sector"
- "[Country] public financial management PFM reform budget execution"
- "[Country] tax revenue GDP fiscal consolidation"
- "[Country] tax to GDP ratio tax burden comparison"

### Step 6 — Labor market in depth

Never leave the labor section as a one-line "macro good, micro bad".

- "[Country] labor market informality underemployment"
- "[Country] youth unemployment sectoral employment"
- "[Country] real wages productivity labor share"
- "[Country] employment growth by sector"

### Step 7 — Policy and reform context

Grounds the recommendations section — do not write recommendations without it.

- "[Country] fiscal consolidation tax reform recommendations"
- "[Country] debt management strategy IMF World Bank"
- "[Country] structural reform priorities"
- "[Country] demographic dividend youth bulge policy"

### Step 8 — Market implications and FDI

- "[Country] equity market outlook"
- "[Country] bond market yields spreads"
- "[Country] currency forex outlook"
- "[Country] FDI foreign direct investment inflows outflows"
- "[Country] FDI trajectory outlook greenfield M&A"

### Step 9 — Regional context (if applicable)

- "[Country] vs peers economic performance"
- "G7 economic comparison GDP inflation rates"
- "developed markets emerging markets outlook"

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

This deliverable is written for institutional, multilateral, and academic readers. Non-negotiables:

- **Structural and historical context** present — sector transformation over time, labor productivity by sector
- **Debt and PFM mechanics** quantified: composition, servicing burden, tax-to-GDP versus peers or IMF/OECD benchmarks
- **Labor market** covered at both macro and micro level (informality, underemployment, youth, real wages vs productivity)
- **Dedicated policy recommendations section**, structured by pillar — fiscal consolidation, debt management, monetary policy, structural reforms — with specific targets and mechanisms where available, each sourced
- No shallow, point-in-time-only narrative

Referenced files: 4

View saved version →

---
name: bigdata-country-sector-analysis
description: >
  Analyze a specific sector inside a specific country or region using Bigdata.com data —
  combining the macroeconomic backdrop (GDP, inflation, rates, policy), country-specific sector
  trends and valuations, fundamentals of country-domiciled sector leaders confirmed by
  geographic revenue exposure, the policy and regulatory environment including subsidies,
  tariffs and foreign-investment rules, and valuation versus global sector peers. Use whenever
  a request names BOTH a sector AND a country or region. Triggers: "macro analysis of X in Y",
  "European financials outlook", "US technology sector view", "India consumer sector",
  "China EV sector", "Japanese semiconductor industry", "[sector] in [country]".
---

# Bigdata Country-Sector Analysis

Sector view anchored to one country or region. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the request combines **both** a sector and a country/region — "Technology investment in the USA", "European financials outlook", "India consumer sector", "China EV". Not this skill when:

| Request | Use instead |
|---------|-------------|
| A sector globally, with no country anchor | Sector analysis |
| A country's economy, with no sector anchor | Country analysis |
| Several sectors compared | Cross-sector comparison |
| Several regions compared | Regional comparison |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_search` | Country macro backdrop, country-sector trends, policy | None |
| `find_securities` | Entity ids for country-domiciled sector leaders | None |
| `bigdata_company_tearsheet` | Fundamentals and geographic revenue split | `find_securities` |
| `bigdata_events_calendar` | Upcoming earnings for those leaders | `find_securities` |
| `bigdata_country_tearsheet` | Economic data where available | None |

## Workflow

### Step 1 — Country economic context

- "[Country] economic outlook GDP growth 2026"
- "[Country] inflation interest rates monetary policy"
- "[Country] central bank rate decision outlook"

Extract GDP growth, inflation, interest rates, and the policy environment.

### Step 2 — Country-specific sector news

Query **country + sector** together:

- "[Country] [Sector] sector outlook 2026"
- "[Country] [Sector] industry trends performance"
- "[Country] [Sector] regulatory policy government"
- "[Country] [Sector] investment flows foreign domestic"
- "[Country] [Sector] earnings revenue growth"
- "[Country] [Sector] valuations multiples"
- "[Country] [Sector] headwinds risks challenges"
- "[Country] [Sector] tailwinds opportunities growth drivers"

### Step 3 — Country-domiciled sector leaders

Use `find_securities` for 5–10 companies **headquartered in** or **primarily operating in** that country, then `bigdata_company_tearsheet` for each:

- Revenue breakdown by geography — **confirm the country exposure is real**, not just a listing venue
- Financial metrics and performance
- Analyst estimates and sentiment
- Hiring trends as a workforce signal

### Step 4 — Events

Use `bigdata_events_calendar` for upcoming earnings across those leaders; filter by the country's exchange for a market-wide scan.

### Step 5 — Policy and regulation

- "[Country] [Sector] regulation policy 2026"
- "[Country] government [Sector] subsidies incentives"
- "[Country] [Sector] trade tariffs exports"
- "[Country] [Sector] foreign investment restrictions"

### Step 6 — Synthesize the macro-sector view

Combine the country backdrop, the country-specific sector trends, company fundamentals, the policy environment, and **valuation relative to global sector peers**. The value of this deliverable is the intersection — a generic sector view with a country label pasted on top does not pass.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

Non-negotiables:

- Country macro backdrop **and** sector detail both present — neither alone is the deliverable
- Company selection justified by domicile or operations, with geographic revenue confirming it
- Policy, subsidies, tariffs, and foreign-investment rules addressed — these usually dominate country-sector outcomes
- Valuation framed against **global** sector peers, so the country premium or discount is visible
- Every claim from a source carries an inline citation and appears in the Sources table

Referenced files: 4

View saved version →

---
name: bigdata-cross-sector
description: >
  Compare two or more sectors using Bigdata.com data — relative valuations, earnings growth,
  analyst sentiment, and where each sits in the economic cycle — and turn that into a rotation
  call with overweight and underweight recommendations. Includes bellwether-level fundamentals
  per sector and a profitability/ROIC-versus-history read that says whether current valuations
  sit on peak, mid-cycle, or trough earnings power. Triggers: "compare X vs Y sectors",
  "which sectors look attractive", "sector rotation", "cyclicals vs defensives",
  "relative value across sectors", "should I rotate out of X into Y".
---

# Bigdata Cross-Sector Comparison

Relative value and rotation across sectors. Use Bigdata.com plugin tools for every fact.

**Use this skill when** two or more sectors are being weighed against each other. Not this skill when:

| Request | Use instead |
|---------|-------------|
| One sector in depth | Sector analysis |
| A sector inside one country | Country-sector analysis |
| An actionable playbook for investing one sector | Sector playbook |
| Regions rather than sectors | Regional comparison |
| Individual companies within a sector | Peer comparables |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_search` | Sector performance, valuation, growth, cycle context | None |
| `find_securities` | Entity ids for 3–5 bellwethers per sector | None |
| `bigdata_company_tearsheet` | Bellwether fundamentals and estimates | `find_securities` |

## Workflow

### Step 1 — Define the sectors in scope

GICS sectors: Information Technology, Health Care, Financials, Consumer Discretionary, Consumer Staples, Industrials, Energy, Materials, Real Estate, Communication Services, Utilities. If the user named sectors, use theirs; otherwise confirm which to compare rather than sweeping all eleven.

### Step 2 — Gather sector data

For **each** sector in scope:

- "[Sector] sector performance valuation"
- "[Sector] sector earnings growth estimates"
- "[Sector] sector analyst recommendations"

### Step 3 — Select bellwethers

Use `find_securities` for 3–5 companies per sector, then `bigdata_company_tearsheet` for each. These anchor the sector-level numbers in something checkable.

### Step 4 — Economic cycle analysis

- "sector rotation economic cycle"
- "cyclical vs defensive outlook"
- "interest rate sensitive sectors"

### Step 5 — Profitability and ROIC spread context

For **each** sector, add a short read on profitability versus history (or versus cost of capital), using bellwether tearsheets and search:

- "[Sector] sector ROIC margin cycle vs historical average"
- "sector profitability peak trough"

State whether current valuations sit on **peak**, **mid-cycle**, or **trough-like** earnings power — where the evidence allows. This is the difference between a comparison that misleads and one that informs: a low P/E on peak earnings is not cheap. Deeper framework: [references/porter-five-forces.md](./references/porter-five-forces.md).

### Step 6 — Rotation call

Rank the sectors and state the rotation explicitly: what to overweight, what to underweight, and the specific reason for each. Tie the call to cycle positioning and the earnings-power read, not to trailing multiples alone.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

Non-negotiables:

- Every sector in scope covered on the **same** metrics, so the comparison is like-for-like
- Cycle positioning stated per sector, not just aggregate market commentary
- Peak / mid / trough earnings-power read attempted, or its data limits flagged
- A rotation call actually made — overweight and underweight, with reasons
- Every claim from a source carries an inline citation and appears in the Sources table

Referenced files: 5

View saved version →

---
name: bigdata-earnings-digest
description: >
  Analyze a public company's latest reported earnings — a cited post-print digest using
  Bigdata.com data (results, consensus and surprise, transcript, analyst reactions, tearsheet
  financials). Breaks down revenue and margins, segment and operating KPIs, management guidance,
  cash flow and balance sheet, and surprises versus expectations with sustainable-vs-one-time
  framing, plus a bull/bear thesis check, quality signals with forward watch-fors, sentiment and
  positioning, a post-print scenario refresh with probability-weighted expected value, and a
  valuation cross-check. Triggers: "analyze X earnings", "earnings digest for X", "how did X
  do last quarter", "X Q3 results", "break down X's earnings", "what did X report",
  "post-earnings analysis", "did X beat or miss".
---

# Bigdata Earnings Digest

Deep dive on one earnings event that has already been reported. Use Bigdata.com plugin tools for every fact.

**Use this skill when** results are out and the user wants them broken down. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Analysis ahead of an upcoming print | Earnings preview |
| 30 days of all developments, not one event | Company brief |
| "What is it worth" with no earnings event | Valuation snapshot |
| Comprehensive risk mapping | Risk assessment |
| Short reaction note against a stated thesis | Earnings reaction |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Latest quarter, consensus, surprise, segments, history, sentiment/positioning fields | `find_securities` |
| `bigdata_events_calendar` | Date of the most recent earnings call | `find_securities` |
| `bigdata_search` | Release, transcript, analyst reactions, guidance, legal coverage | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Before you synthesize — quality over quantity

The digest is **forward-looking**, not a transcription of the release. Before writing, identify the **2–3 factors that dominate the forward debate after this print** and lead with them. Everything else is supporting detail — do not give every line item equal weight.

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Financial data

Call `bigdata_company_tearsheet` with the `entity_id` for:

- Latest quarterly results (most recent Q)
- Analyst estimates and consensus
- Latest earnings surprise data
- Historical trends for comparison
- Segment performance breakdown
- **Sentiment, ownership, insider, options/short** fields where exposed — these fill the structured positioning table

If more than one recent quarter is available, confirm which to analyze:

> "I see results for Q3 2024 (most recent) and Q2 2024. Should I analyze the latest Q3 results?"

### Step 3 — Earnings date

Call `bigdata_events_calendar` with the `entity_id` to pin down when the most recent earnings call was.

### Step 4 — Search earnings materials

Run **5–7 targeted `bigdata_search` queries**:

- "[Company] earnings results Q[X] [Fiscal Year]"
- "[Company] earnings transcript conference call"
- "[Company] analyst reactions upgrades downgrades"
- "[Company] guidance outlook management commentary"
- "[Company] earnings surprise beat miss"
- "[Company] lawsuit litigation regulatory ruling investigation" — post-print legal overhang not in the press release

Cover the official release and metrics, transcript highlights, analyst reactions and rating changes, management guidance, and market reaction. Extract **key quotes** from the transcript where available.

If the call isn't published yet:

> "The earnings call transcript isn't available yet. I'll analyze the press release and update once the call is published."

### Step 5 — Results analysis

Organize the numbers into:

- **Revenue and margins** — total vs consensus, by segment/geography, gross/operating/net margin, YoY and QoQ
- **Operating metrics and segments** — KPIs, segment results, customer/user metrics, geographic performance
- **Management guidance and commentary** — forward guidance vs consensus, strategic initiatives, market conditions, capital allocation
- **Cash flow and balance sheet** — OCF and FCF trends, balance sheet strength, capex and investments

Note accounting changes and one-time items explicitly.

### Step 6 — Surprises: magnitude and quality

For each beat or miss:

- Quantify in **% or bps** vs consensus
- Frame **magnitude** as approximate **standard deviations** against the company's typical surprise volatility where data allows
- Label it **sustainable vs one-time** — revenue volume or price vs buyback, tax, timing, or other one-timers
- Identify what actually drove the market reaction

Focus on business fundamentals, not just the stock move.

### Step 7 — Thesis check (forward-looking)

Even with no thesis supplied by the user, frame both sides:

- **For bulls:** this quarter **strengthened / weakened / left unchanged** the bull case because [specific evidence].
- **For bears:** this quarter **strengthened / weakened / left unchanged** the bear case because [specific evidence].

If the user *did* supply a thesis, state its status explicitly as **Intact / Strengthened / Weakened / Broken**, with the specific data points that support the call. Methodology: [references/thesis-construction.md](./references/thesis-construction.md).

### Step 8 — Quality signals

Build the table with a forward **watch for** column — monitoring, not just a backward check:

| Signal | This quarter | Prior quarter | Trend / note | **Watch for (forward)** |
|--------|--------------|---------------|--------------|-------------------------|
| OCF vs net income | | | | |
| DSO | | | | |
| Inventory (if material) | | | | |
| Guidance vs actual (credibility) | | | | |

Depth: [references/quality-of-earnings.md](./references/quality-of-earnings.md).

### Step 9 — Sentiment & positioning (structured, not anecdotes)

Same discipline as the earnings preview: pull every **numeric** sentiment, insider, 13F/flow, and options/short field from the tearsheet first, then use Step 4 results to fill gaps. Write **"Not available"** for missing cells rather than dropping rows. A single sell-side note is not a substitute for positioning data.

### Step 10 — Scenario refresh + valuation cross-check

**Scenario refresh (post-print):** rebuild Bull / Base / Bear against the *new* information — probabilities summing to ~100%, what changed versus pre-print, price level or range, and the **probability-weighted expected value with the arithmetic shown**. Compute in prose/table by default; run [scripts/scenario_probability.py](./scripts/scenario_probability.py) only if the user explicitly asks for scripted math.

**Valuation cross-check:** current EV/EBITDA, P/E, FCF yield from the tearsheet vs recent history and peers. Answer directly: **does the reaction fit the surprise and the guidance?** Does the price now embed the new guidance?

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, mandatory tables, sources, and footer.

- Add inline citations as superscript-style numbers `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version at the end if useful.

## Quality bar

Pass the PM test before delivering: **What's different?** **What matters (2–3 forward factors)?** **What should I do about it?** (net assessment, key risk, next catalyst — no position sizing). Would it survive a short, skeptical morning meeting without reading as a press-release recap?

Non-negotiables in every digest:

- Thesis check for **both** bull and bear, with evidence — this is what makes the digest forward-looking
- Surprise magnitude quantified, and labelled **sustainable vs one-time**
- Quality signals table with the **watch for** column filled
- Sentiment & positioning as structured data — tearsheet first, then search
- Scenario refresh with probabilities and **EV math shown**
- Valuation cross-check tying implications back to price
- Legal/regulatory search run, with overhangs surfaced
- Facts separated from analysis and implications

Referenced files: 7

View saved version →

---
name: bigdata-earnings-preview
description: >
  Create a forward-looking earnings preview for a public company ahead of its next earnings
  call, using Bigdata.com MCP data (estimates, tearsheet financials, news, filings, transcripts,
  events calendar). Produces an EPIC driver table, earnings quality screen with forward
  watch-fors, structured sentiment and positioning data, what's priced in plus valuation
  cross-check, FaVeS variant perception, bull and bear cases, bull/base/bear scenarios with
  probability-weighted expected value, and key metrics to watch — fully cited. Triggers:
  "earnings preview for X", "preview X earnings", "pre-earnings analysis", "Q3 preview",
  "what to expect before X reports", "what should I watch when X reports", "set up for X
  earnings", "bull and bear case into the print".
---

# Bigdata Earnings Preview

Forward-looking, pre-earnings research note on a public company. Use Bigdata.com plugin tools for every fact; apply the pre-synthesis filter below before writing.

**Use this skill when** the user wants analysis *before* a company reports. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Analysis of results already reported | Earnings digest / earnings reaction |
| Retrospective summary of recent news | Company brief |
| "What is it worth" with no earnings event | Valuation snapshot |
| Comprehensive risk mapping | Risk assessment |

## Data foundation (MCP tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Financials, estimates, margins, sentiment/positioning fields | `find_securities` |
| `bigdata_events_calendar` | Next earnings call date | `find_securities` |
| `bigdata_search` | News, filings, transcripts, analyst and regulatory coverage | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Before you synthesize — quality over quantity

Do this **after** gathering data and **before** writing the draft:

1. List candidate drivers from tearsheet + search.
2. Rank them and keep the top **2–3** as primary drivers. Do not give 20 findings equal weight.
3. Run the full **EPIC** filter on each primary driver:

| Test | Question |
|------|----------|
| **E**ffect (material) | Would getting this wrong move value or the debate on the name meaningfully? |
| **P**redictability | Can *you* form a view with evidence, not speculation? |
| **I**ndependence | Does consensus or price **systematically** under- or over-weight this factor? |
| **C**onsensus gap | Does **your** view differ from consensus in a specific, falsifiable way? |

4. Map each primary driver to an implication (bullish / bearish / neutral) with specific metrics.
5. Deprioritize template sections that are immaterial this period — say so in one line rather than padding.

Depth: [references/epic-framework.md](./references/epic-framework.md).

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Financial baseline

Call `bigdata_company_tearsheet` with the `entity_id` for:

- Recent quarterly performance trends and YoY comparisons
- Historical earnings surprises
- Analyst estimates for the upcoming quarter
- Key financial metrics and margins
- **Positioning / sentiment fields when exposed** — sentiment scores, news/social metrics, ownership concentration, insider summary, options or short interest. Capture whatever the tool returns; never substitute analyst headlines for systematic data when numbers exist.

### Step 3 — Earnings quality quick screen

Before building narratives, record a credibility table with a forward-looking **watch for** column (approximate if necessary; flag data gaps):

| Check | This period / trend | Red-flag threshold | **Watch for (next print)** |
|-------|---------------------|--------------------|----------------------------|
| OCF / Net income | | Healthy often >0.8 sustained; <0.6 or widening gap → dig | further OCF/NI divergence; one-time boosts rolling off |
| DSO vs revenue growth | | DSO rising faster than revenue → recognition risk | DSO days vs rev growth; channel inventory mentions |
| GAAP vs non-GAAP EPS gap | | Large or widening gap → quality question | stock comp, restructuring, "adjusted" add-backs |

If the tearsheet lacks a line, search: "[Company] operating cash flow vs net income non-GAAP reconciliation".

Depth: [references/quality-of-earnings.md](./references/quality-of-earnings.md).

### Step 4 — Earnings date

Call `bigdata_events_calendar` with the `entity_id` to find the next earnings call. If unknown, ask:

> "I don't have the exact earnings date yet. Shall I proceed with the preview based on recent developments and expectations?"

### Step 5 — Search: developments, legal/regulatory, positioning

Cast a **wide net** so material non-operational risks (court rulings, probes, tax disputes) are not missed. Use `bigdata_search` over the last **60–90 days** (extend if coverage is thin). Run **at least 8–10 targeted queries** across all three buckets, then merge redundant results.

**Core company & industry**
- "[Company] recent developments last 90 days"
- "[Company] product launches initiatives"
- "[Company] guidance commentary management"
- "[Company] analyst expectations earnings preview"
- "[Industry] trends headwinds tailwinds"

**Regulatory, legal, policy (mandatory)**
- "[Company] lawsuit litigation court ruling settlement regulatory investigation last 90 days"
- "[Company] SEC investigation DOJ antitrust fine penalty Europe"
- "[Company] tax dispute regulatory approval compliance"

**Market positioning & flows (mandatory — fills the structured table in the output)**
- "[Company] insider buying selling Form 4 transactions last 90 days"
- "[Company] institutional ownership 13F changes fund flows"
- "[Company] short interest options put call ratio open interest"
- "[Company] news sentiment score" (or closest available)

If news is sparse, ask:

> "There's been limited news recently. Would you like me to expand the search period or focus on industry trends?"

### Step 6 — Sentiment & positioning (structured, not anecdotes)

Build the output table from data, not from a single analyst note:

1. Pull every **numeric** sentiment / flow / positioning field from the tearsheet.
2. Use Step 5 results to fill gaps (insider trades, large holder moves, options/skew, quantified sentiment).
3. If a cell is unavailable, write **"Not available in data"** — the section still appears.

### Step 7 — What's priced in + valuation cross-check

**Before** writing bull/bear narratives, establish what the current price embeds for this quarter and the near-term trajectory. Use tearsheet multiples, consensus, and reverse-DCF-style reasoning (conceptual is fine — [references/reverse-dcf.md](./references/reverse-dcf.md)).

| Lens | Implied by market | Consensus | Your assessment |
|------|-------------------|-----------|-----------------|
| Growth (revenue / key volume) | | | |
| Margin level or expansion | | | |
| Beat magnitude / "whisper" vs published consensus | | | |

**Multiples sanity check:** current EV/EBITDA, P/E, FCF yield (or sector-standard multiples) vs ~5-year range or peer median where data allows. State whether valuation implies **optimism**, **consensus**, or **pessimism** relative to the setup.

### Step 8 — Variant perception (FaVeS) + scenarios

**FaVeS — mandatory structure in the output:**
- **Fundamentals** — the 2–3 KPIs that drive the quarter; where consensus could be wrong (link to bull/bear).
- **Valuation** — tie to the *What's priced in* table and valuation cross-check; cross-reference rather than repeat prose.
- **Sentiment** — tie to the *Sentiment & positioning* table; separate what is priced in **behaviorally** from **fundamentally**.

Depth: [references/faves-framework.md](./references/faves-framework.md).

**Scenario analysis — mandatory:** build Bull / Base / Bear with

- **Probability weights** summing to ~100% (e.g. 30/50/20), briefly justified
- **Key assumptions** per scenario (growth, margin, one-timers, legal outcomes)
- **Price level or range** per scenario (spot, consensus PT band, or a rough DCF/multiple bridge — show the assumptions)
- **Probability-weighted expected value** with the arithmetic shown (EV = Σ p×P; state expected upside/downside % vs spot)

Methodology: [references/thesis-construction.md](./references/thesis-construction.md). Compute in prose/table by default; run [scripts/scenario_probability.py](./scripts/scenario_probability.py) only if the user explicitly asks for scripted math.

### Step 9 — Synthesize

Lead with the 2–3 primary drivers and their EPIC documentation. Cover, in order of materiality only:

- **Recent developments** — launches, partnerships or M&A, operational shifts, geographic/share changes, plus material legal and regulatory items
- **Industry trends** — macro drivers, competitive landscape, supply chain and cost pressure, policy
- **Bull case** — each point specific, measurable, evidence-backed and resolvable over a sensible horizon; tie to consensus line items where possible ("consensus models X% growth in segment Y; channel evidence suggests Z%, ~$Nm revenue upside"); cite a source per claim
- **Bear case** — same discipline; quantify downside (margin bps, revenue %, one-time vs recurring)
- **Key metrics to watch** — the KPIs that matter *this* quarter, those that move the stock given what's priced in, and where surprise volatility is highest vs whisper/consensus

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, mandatory tables, sources, and footer.

- Add inline citations as superscript-style numbers `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version at the end if useful.

## Quality bar

Pass the PM test before delivering: **What's different?** **What matters (2–3 drivers)?** **What should I do about it?** (net assessment, key risk, next catalyst — no position sizing). Would it survive a short, skeptical morning meeting without reading as a data dump?

Non-negotiables in every preview:

- EPIC table for each elevated driver — filled, not placeholder
- Scenario table with probabilities, prices, and **EV math shown**
- Sentiment & positioning as structured data — tearsheet first, then search
- Regulatory/legal queries run, and material items surfaced in developments and the bear case
- **Watch for** column on the quality screen — forward monitoring, not only backward checks
- *What's priced in* built before bull/bear, so both cases are relative to embedded expectations
- Facts separated from analysis and implications

Referenced files: 11

View saved version →

---
name: bigdata-earnings-quality-screen
description: >
  Screen a public company's reported earnings for quality and accounting red flags using
  Bigdata.com data and filings. Covers cash conversion (OCF/NI, FCF/NI across periods), accruals
  and the balance-sheet accrual ratio, working-capital signals (DSO, DIO, DPO versus revenue
  growth), revenue-recognition and capitalization flags, the GAAP versus non-GAAP gap and the
  nature of the add-backs, and an optional Beneish M-Score with inputs shown — closing with a
  verdict on how far the reported numbers can be trusted. Triggers: "earnings quality screen
  for X", "are X's earnings real", "accounting red flags at X", "is X manipulating earnings",
  "cash conversion at X", "check X's accruals", "quality of earnings on X".
---

# Bigdata Earnings Quality Screen

Forensic check on whether reported earnings are backed by cash. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the question is whether the numbers can be trusted. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Full breakdown of a reported quarter | Earnings digest |
| All risk categories, rated | Risk assessment |
| Valuation of the business | Valuation snapshot |
| Full thesis with recommendation | Investment memo |

A quality screen is **diagnostic, not accusatory**. Aggressive accounting is common and often legal; the deliverable is a graded read on reliability, with the evidence shown, not an allegation.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Income statement, cash flow, balance sheet across periods | `find_securities` |
| `bigdata_search` | Filings, reconciliations, restatements, auditor and short-seller commentary | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Pull multi-period data

Call `bigdata_company_tearsheet`. **Trend is the signal** — a single period tells you almost nothing. Get at least 4–8 quarters, or 3 years, of net income, operating cash flow, free cash flow, receivables, inventory, payables, revenue, and total assets.

### Step 3 — Cash conversion

| Check | Healthy | Investigate |
|-------|---------|-------------|
| OCF / Net income | >0.8 sustained | <0.6, or a widening gap over time |
| FCF / Net income | Positive and stable | Persistently negative while NI is positive |

A company that reports profits but does not generate cash is the single most common quality problem. Chart the trend, don't just take the latest ratio.

### Step 4 — Accruals

- **Accrual ratio (balance sheet)** = (net operating assets end − net operating assets start) / average net operating assets
- **Accrual ratio (cash flow)** = (net income − OCF − investing cash flow) / average total assets

High and rising accruals mean earnings are increasingly made of estimates rather than cash. Show the inputs.

### Step 5 — Working capital signals

| Signal | Red flag |
|--------|----------|
| DSO vs revenue growth | Receivables growing faster than revenue → recognition or collection risk |
| DIO / inventory | Inventory building ahead of sales → demand weakness or write-down risk |
| DPO | Stretching payables → liquidity strain dressed as cash flow |

### Step 6 — Revenue recognition and capitalization

Search for the specifics:

- "[Company] revenue recognition policy change"
- "[Company] capitalized software development costs"
- "[Company] restatement auditor change material weakness"
- "[Company] related party transactions revenue"

Look for: recognition timing changes, capitalizing what peers expense, revenue from related parties, channel stuffing signals, and unusual "other income".

### Step 7 — GAAP versus non-GAAP

Size the gap and — more importantly — characterize the add-backs. Recurring "one-time" restructuring, perpetual stock-comp exclusion, and adjustments that only ever go one direction are the tell. Search: "[Company] non-GAAP reconciliation adjusted EBITDA add-backs".

### Step 8 — Optional Beneish M-Score

When several signals above are flashing, compute the Beneish M-Score and show the eight inputs (DSRI, GMI, AQI, SGI, DEPI, SGAI, LVGI, TATA). Flag data gaps rather than guessing inputs. Run [scripts/earnings_quality.py](./scripts/earnings_quality.py) only if the user wants scripted output; otherwise compute in the table.

Frameworks: [references/quality-of-earnings.md](./references/quality-of-earnings.md), [references/red-flags-checklist.md](./references/red-flags-checklist.md).

### Step 9 — Verdict

Grade the overall quality — **High / Adequate / Questionable / Poor** — and state the specific evidence behind the grade, plus what would confirm or clear each concern in the next print.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) version if useful.

## Quality bar

Non-negotiables:

- **Multi-period trends**, not single-period ratios
- Every flag carries the arithmetic or the source that produced it
- Data gaps stated explicitly — never fill a missing input with a guess
- Add-backs characterized, not just totalled
- A graded verdict given, with what would clear each concern
- Diagnostic language throughout — evidence and probability, not accusation

Referenced files: 7

View saved version →

---
name: bigdata-earnings-reaction
description: >
  Write a tight post-earnings reaction note using Bigdata.com data — headline numbers versus
  consensus with beat/miss magnitude, what mattered on both sides, a prior-versus-new guidance
  table, an explicit thesis check (Intact / Strengthened / Weakened / Broken) with evidence,
  the estimate and price-target revisions the print forces, a pre-versus-post valuation update,
  quality signals for the quarter, and an action with the next key date. Shorter and more
  decision-focused than a full earnings digest. Triggers: "earnings reaction for X", "how
  should I react to X's results", "does X's quarter change the thesis", "X print reaction",
  "revise my numbers after X earnings", "was X's quarter good enough".
---

# Bigdata Earnings Reaction

The decision note after a print: what changed, what to revise, what to do. Use Bigdata.com plugin tools for every fact.

**Use this skill when** results are out and the user needs a fast, thesis-anchored call. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Full breakdown of the quarter — segments, cash flow, guidance detail | Earnings digest |
| Analysis ahead of the print | Earnings preview |
| A view with no earnings event | Quick take |
| Full thesis rebuild | Investment memo |

**Digest vs reaction:** the digest explains the quarter; the reaction decides what to do about it. If the user has a position or a stated thesis, this is the right skill.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Reported numbers, consensus, surprise, multiples pre/post | `find_securities` |
| `bigdata_events_calendar` | Confirm the report date and the next key date | `find_securities` |
| `bigdata_search` | Release, transcript, guidance, analyst reactions | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Anchor the print

Resolve the entity, confirm the reported date, and pull the tearsheet for actuals versus consensus, the surprise, and multiples before and after the move.

### Step 2 — Ask for the thesis (or infer both sides)

If the user has stated a thesis, use it — the thesis check is the centerpiece of this note. If not, construct the prevailing bull and bear cases from the tearsheet and search, and check the quarter against **both**.

### Step 3 — Headline numbers

Revenue, EPS, and the key sector KPI: reported, consensus, beat/miss in % or bps, and YoY change. Quantify — "solid quarter" is not a number.

### Step 4 — What mattered

Positive and negative surprises, each with its magnitude and whether it is **sustainable or one-time** (revenue volume or price vs buyback, tax, timing). Run 3–5 searches:

- "[Company] earnings results Q[X] [Fiscal Year]"
- "[Company] earnings call transcript guidance commentary"
- "[Company] analyst reactions price target changes"
- "[Company] earnings surprise beat miss reaction"

### Step 5 — Guidance update

Prior guidance versus new guidance versus consensus, per metric. Guidance usually moves the stock more than the reported quarter — treat it as first-order.

### Step 6 — Thesis check

State the status explicitly: **Intact / Strengthened / Weakened / Broken**. Back it with the specific data points from the quarter that support the call — not a general impression.

### Step 7 — Revisions and valuation

What the print forces you to change: FY revenue, FY EPS, price target. Then the valuation update — stock price, NTM P/E, NTM EV/EBITDA, pre- versus post-earnings. Does the move fit the news?

### Step 8 — Quality signals

OCF vs net income, DSO trend, inventory build, and guidance credibility (met/beat versus missed). A beat on declining quality is a different result than a beat on improving quality. Depth: [references/quality-of-earnings.md](./references/quality-of-earnings.md).

### Step 9 — Action

Give the action and the rationale in one or two sentences, plus the next key date. If scenarios need re-weighting, run [scripts/scenario_probability.py](./scripts/scenario_probability.py) only when the user asks for scripted math.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) version if useful.

## Quality bar

Non-negotiables:

- Thesis status stated as one of the four words, with evidence — the whole note turns on this
- Beat/miss quantified and labelled sustainable vs one-time
- Guidance treated as first-order, with prior versus new side by side
- Estimate and price-target revisions named explicitly, not left implied
- An action given with a next key date
- Facts separated from analysis and implications

Referenced files: 6

View saved version →

---
name: bigdata-g7-comparison
description: >
  Benchmark the seven G7 economies side by side using Bigdata.com data — the United States,
  Japan, Germany, the United Kingdom, France, Italy, and Canada. Produces a like-for-like
  indicator table (GDP, inflation, unemployment, policy rate, fiscal position), central bank
  stance and rate-path divergence across the Fed, BoJ, ECB, BoE and BoC, relative market
  positioning across equities, rates and currencies, the divergence and convergence themes
  running through the bloc, and a ranked allocation view — optionally focused on equities,
  rates, FX or credit. Triggers: "compare G7 economies", "G7 comparison", "G7 outlook",
  "how do the G7 economies compare", "G7 growth and inflation", "G7 central bank divergence".
---

# Bigdata G7 Comparison

Side-by-side benchmark of the seven G7 economies. Use Bigdata.com plugin tools for every fact.

**Members:** United States, Japan, Germany, United Kingdom, France, Italy, Canada. Note the euro-area members (Germany, France, Italy) share the ECB's policy rate — differentiate them on fiscal position, growth, and market pricing rather than on monetary policy.

**Use this skill when** the G7 is the frame. Not this skill when:

| Request | Use instead |
|---------|-------------|
| One G7 country in depth | Country analysis |
| Broader or different regions (EM, Asia ex-Japan) | Regional comparison |
| Sectors rather than economies | Cross-sector comparison |
| A macro theme across borders | Thematic research |

If the user supplied a focus (equities, rates, FX, credit), lead the market-implications section with it. Otherwise cover all four at a broad level.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_country_tearsheet` | Economic data and built-in G7 comparison where available | None |
| `bigdata_search` | Indicators, policy, market positioning per member | None |

If `bigdata_country_tearsheet` is unavailable or fails, complete the analysis with `bigdata_search` alone.

## Workflow

### Step 1 — Indicators for each member

Pull the **same** indicators for all seven so the table is like-for-like: GDP growth, CPI inflation, unemployment, policy rate, and fiscal position (deficit and debt/GDP).

- "[Country] GDP growth economic outlook 2026"
- "[Country] inflation CPI consumer prices"
- "[Country] unemployment labor market"
- "[Country] government deficit debt to GDP"

Where a figure is unavailable for one member, write "Not available" rather than leaving the comparison lopsided.

### Step 2 — Central bank stance and rate paths

- "Fed ECB BOJ BOE rate decision outlook"
- "[Country] central bank policy rate path expectations"
- "G7 monetary policy divergence"

Capture the last action, current guidance, and market-implied path for the **Fed, BoJ, ECB, BoE, and BoC**. Policy divergence across the bloc is usually the single most important driver of relative returns — treat it as a headline finding, not a footnote.

### Step 3 — Comparative economic analysis

- "G7 economic comparison GDP growth rates"
- "G7 inflation comparison"
- "G7 fiscal position debt sustainability"

Identify who is leading and lagging on growth, who is winning and losing the inflation fight, and whose fiscal position constrains policy.

### Step 4 — Market positioning

- "G7 equity market valuations comparison"
- "G7 sovereign bond yields comparison"
- "USD EUR JPY GBP CAD currency outlook"
- "credit spreads investment grade high yield [region]"

Cover equity valuations, 10-year yields and curve shape, currency levels and outlook, and credit where relevant.

### Step 5 — Divergence and convergence themes

Name the 2–3 themes running through the bloc — for example policy divergence, fiscal stress, energy exposure, demographics, or trade policy — and which members each helps or hurts.

### Step 6 — Ranked view

Rank the seven on relative attractiveness for the user's focus (or overall), with a one-line reason each and the key risk to the ranking.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

Non-negotiables:

- All seven members present in the indicator table, on the **same** metrics, with gaps marked "Not available"
- Euro-area members differentiated on fiscal position, growth, and market pricing — not treated as one country
- Central bank divergence treated as a headline driver
- A ranked view delivered, with reasons and the key risk
- Every claim from a source carries an inline citation and appears in the Sources table

Referenced files: 4

View saved version →

---
name: bigdata-investment-memo
description: >
  Write a full institutional investment memo on a public company using Bigdata.com data —
  thesis, variant perception versus consensus, valuation, risks, catalysts, and an explicit
  recommendation with conviction. Runs the complete workflow: EPIC-filtered primary drivers,
  FaVeS variant perception, earnings quality and moat assessment, valuation by the method that
  fits the business with a secondary cross-check, bull/base/bear scenarios with probabilities
  and a probability-weighted value, key risks and what would change the view. Triggers:
  "investment memo for X", "full analysis of X", "should I buy X", "build the bull case for X",
  "write up X as an investment", "deep dive on X", "thesis on X", "DCF thesis for X".
---

# Bigdata Investment Memo

The deepest company deliverable: a complete thesis with an explicit recommendation. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants the full write-up. Not this skill when:

| Request | Use instead |
|---------|-------------|
| A fast view, one page | Quick take |
| Just "what is it worth" | Valuation snapshot |
| Just the consensus gap | Variant perception |
| Just bull/base/bear with probabilities | Scenario analysis |
| Just risks, rated | Risk assessment |
| Recent developments | Company brief |

## Core philosophy

Anchor everything on three ideas:

1. **Intrinsic value** — estimate what the business is worth independent of the price
2. **Variant perception** — state clearly where your view differs from consensus
3. **Quality over quantity** — prioritize the few drivers that matter, do not weight twenty findings equally

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Financials, estimates, margins, sentiment, segments | `find_securities` |
| `bigdata_search` | News, filings, transcripts, analyst and competitive coverage | None |
| `bigdata_events_calendar` | Upcoming earnings and conferences | `find_securities` |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Company and data

Resolve the entity, pull the tearsheet, and run focused searches for recent developments, competitive position, management commentary, and the current debate on the name. Establish the factual base before any analysis.

### Step 2 — What matters (EPIC)

Filter candidate drivers down to the **2–3 that pass all four tests**:

| Test | Question | Pass criteria |
|------|----------|---------------|
| **E**ffect | Is it material? | ~10% change moves intrinsic value meaningfully (e.g. >5%) |
| **P**redictability | Can you forecast it? | You have an analytical or informational edge |
| **I**ndependence | Does consensus get it wrong? | The market systematically misjudges this |
| **C**onsensus gap | Is there a gap? | Your forecast differs meaningfully |

Detail: [references/epic-framework.md](./references/epic-framework.md).

### Step 3 — Variant perception (FaVeS)

| Element | Key questions |
|---------|---------------|
| **Fundamentals** | Which 2–3 KPIs drive value? Where could estimates be wrong? |
| **Valuation** | What is intrinsic value? What multiple fits this quality and growth? |
| **Sentiment** | What is priced in (reverse DCF)? How are investors positioned? |

You **must** articulate where you differ from consensus. Detail: [references/faves-framework.md](./references/faves-framework.md), [references/reverse-dcf.md](./references/reverse-dcf.md).

### Step 4 — Quality and risk (before valuation)

- **Earnings quality:** OCF/NI (healthy typically >0.8; red flag <0.6 or diverging), accruals, DSO vs revenue trend — [references/quality-of-earnings.md](./references/quality-of-earnings.md)
- **Competitive position:** moat type and strength, ROIC vs WACC, competitive advantage period — [references/moat-taxonomy.md](./references/moat-taxonomy.md)
- **Management:** capital allocation, insider activity, guidance track record — [references/capital-allocation.md](./references/capital-allocation.md)

Valuing a business before checking whether its earnings are real is how memos go wrong. Do this step first.

### Step 5 — Value it

Pick the primary method by business type, and always run a secondary check:

| Company type | Primary | Secondary check |
|--------------|---------|-----------------|
| Stable, profitable | DCF (FCFF) | EV/EBITDA, P/E |
| High-growth, pre-profit | EV/Revenue; DCF with long CAP | Reverse DCF |
| Bank / insurer | P/TBV; dividend discount | P/E, residual income |
| REIT | NAV; P/AFFO | Implied cap rate |
| Conglomerate | Sum-of-parts | Holdco discount |
| Distressed | Liquidation / recovery | Asset coverage |

Methodology: [references/dcf-methodology.md](./references/dcf-methodology.md), [references/multiples-framework.md](./references/multiples-framework.md), [references/sum-of-parts.md](./references/sum-of-parts.md). Sector-specific lenses: [references/sector-routing.md](./references/sector-routing.md). Foundations: [references/graham-dodd-principles.md](./references/graham-dodd-principles.md).

### Step 6 — Scenarios

Build **bull / base / bear** with explicit assumptions, probability weights, and a value per scenario. Show the probability-weighted value and the arithmetic. Methodology: [references/thesis-construction.md](./references/thesis-construction.md).

### Step 7 — Risks, catalysts, recommendation

State the key risks and **what would change the view** (falsifiable, not decorative). List dated catalysts. Then give the recommendation and a conviction level — a memo that hedges everything has not done its job.

## Optional scripts

**Default:** work from tearsheet, search, and reasoning — including reverse-DCF reasoning — without running Python.

Use these only when the user explicitly wants spreadsheet-style model output:

| Script | Purpose |
|--------|---------|
| [scripts/dcf_model.py](./scripts/dcf_model.py) | DCF with scenarios |
| [scripts/reverse_dcf.py](./scripts/reverse_dcf.py) | Implied growth extraction |
| [scripts/earnings_quality.py](./scripts/earnings_quality.py) | Beneish M-Score, accruals |
| [scripts/peer_comparables.py](./scripts/peer_comparables.py) | Comp table |
| [scripts/scenario_probability.py](./scripts/scenario_probability.py) | Expected value across scenarios |

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer Word (.docx) for formal memos or a deck-ready structure.

## Quality bar

A memo must clear a concise institutional review. Non-negotiables:

1. Clear **recommendation** and **conviction** (e.g. 1–5)
2. **Explicit variant perception** versus consensus — stated, not implied
3. **Scenarios** with probabilities and price targets or ranges, math shown
4. **Key risks** and what would change the view
5. **Catalysts** with timing
6. Earnings quality and moat assessed **before** the valuation section
7. Facts separated from analysis and implications

Referenced files: 21

View saved version →

---
name: bigdata-moat-governance-review
description: >
  Assess how durable a public company's competitive advantage is and whether management can be
  trusted with the capital, using Bigdata.com data. Covers moat identification by type with
  evidence, moat strength via ROIC versus WACC, pricing power and share trend, a competitive
  advantage period estimate with erosion signals, industry structure via five forces, the
  capital allocation track record across M&A, buybacks, dividends and reinvestment, and
  governance — board independence, dual roles, compensation design, related-party exposure,
  insider activity. Triggers: "does X have a moat", "moat review for X", "how durable is X's
  advantage", "is X's management any good", "capital allocation at X", "governance review of X",
  "competitive advantage of X".
---

# Bigdata Moat & Governance Review

Two questions that decide long-run returns: is the advantage durable, and are the stewards any good? Use Bigdata.com plugin tools for every fact.

**Use this skill when** durability and stewardship are the question. Not this skill when:

| Request | Use instead |
|---------|-------------|
| All risk categories rated | Risk assessment |
| Accounting reliability | Earnings quality screen |
| What the business is worth | Valuation snapshot |
| Full thesis with recommendation | Investment memo |
| Sector-wide structure | Sector analysis / playbook |

These two topics belong together: a wide moat run by poor capital allocators leaks value, and excellent management cannot rescue a business with no structural advantage.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Returns, margins, reinvestment, buybacks, dividends, insider summary | `find_securities` |
| `bigdata_search` | Competitive position, share, pricing, governance, capital allocation history | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company and its industry

Call `find_securities`, then `bigdata_company_tearsheet` for the returns and margin history that the moat assessment rests on.

### Step 2 — Identify the moat by type

Name the moat type — network effects, switching costs, cost advantage, intangibles (brand, patents, licenses), or efficient scale — and give the **evidence** for each claimed source. "Strong brand" without pricing evidence is not a moat finding. Taxonomy: [references/moat-taxonomy.md](./references/moat-taxonomy.md).

Search: "[Company] competitive advantage market share pricing power", "[Company] switching costs customer retention".

### Step 3 — Test moat strength with numbers

| Test | What it shows |
|------|---------------|
| ROIC vs WACC, sustained | Whether the advantage converts to economic profit |
| ROIC trend over 5–10 years | Whether it is widening or eroding |
| Gross and operating margin vs peers | Pricing power in practice |
| Market share trend | Whether the position is being defended |
| Reinvestment rate at high ROIC | Whether the moat has runway |

A moat that does not show up as durable excess returns is a story, not a moat.

### Step 4 — Competitive advantage period and erosion

Estimate how long the advantage plausibly persists, and name the **erosion signals** to monitor: pricing pressure (usually the first sign), share loss to entrants or substitutes, ROIC compression toward WACC, rising customer churn, technology shifts. Industry structure context: [references/porter-five-forces.md](./references/porter-five-forces.md).

### Step 5 — Capital allocation track record

Assess where the cash has gone and what it earned:

- **M&A** — deals done, prices paid, returns achieved, write-downs taken
- **Buybacks** — bought at what valuations; repurchasing above intrinsic value destroys value
- **Dividends** — sustainability against FCF, and consistency
- **Reinvestment** — incremental ROIC on organic capex and R&D
- **Balance sheet** — leverage choices through the cycle

Framework: [references/capital-allocation.md](./references/capital-allocation.md). Search: "[Company] acquisitions track record write-down", "[Company] buyback history capital returns".

### Step 6 — Governance

- Board independence, size, refreshment, and relevant expertise
- Combined CEO/chair role, classified board, dual-class shares and voting concentration
- Compensation design: what metrics vest, over what horizon, and whether they align with per-share value
- Related-party transactions
- Insider buying and selling patterns, read in context rather than mechanically
- Guidance track record — a proxy for candor

Search: "[Company] CEO chairman combined role board independence", "[Company] executive compensation say on pay", "[Company] insider selling Form 4", "[Company] related party transactions".

### Step 7 — Combined verdict

Grade the **moat** (None / Narrow / Wide) and its **trend** (widening / stable / eroding), and grade **management quality** (Strong / Adequate / Weak) on capital allocation and governance separately. Then state what the combination means for the durability of returns, and what would change each grade.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) version if useful.

## Quality bar

Non-negotiables:

- Every claimed moat source backed by **evidence**, not adjectives
- ROIC versus WACC shown over time — the numerical test is mandatory
- Erosion signals named specifically, with what to watch
- Capital allocation judged on **returns achieved**, not on stated intentions
- Governance graded separately from capital allocation — they diverge often
- Both grades come with what would change them

Referenced files: 7

View saved version →

---
name: bigdata-peer-comparables
description: >
  Compare a public company against its peer set using Bigdata.com data — valuation multiples,
  growth, profitability, returns, leverage, and sentiment — to judge relative attractiveness.
  Builds the peer set with an explicit rationale for inclusion and exclusion, tabulates
  like-for-like metrics with peer median and quartile positioning, decomposes any premium or
  discount into what fundamentals justify versus what they do not, and closes with a relative
  verdict. Triggers: "compare X to its peers", "peer comparables for X", "how does X screen vs
  competitors", "is X cheap relative to peers", "comps table for X", "relative valuation of X",
  "who are X's peers".
---

# Bigdata Peer Comparables

Relative screen against a defensible peer set. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the question is relative. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Absolute value — what is it worth | Valuation snapshot |
| Sector-level performance and themes | Sector analysis |
| Sectors ranked against each other | Cross-sector comparison |
| Full thesis with recommendation | Investment memo |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve the subject and every peer → `entity_id` | None |
| `bigdata_company_tearsheet` | Multiples, growth, margins, returns, leverage, sentiment per company | `find_securities` |
| `bigdata_search` | Peer-set validation, competitive positioning, valuation debate | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the subject company

Call `find_securities`, then `bigdata_company_tearsheet` to establish the business model, segment mix, and size.

### Step 2 — Construct the peer set (state the rationale)

Pick **5–8 peers** on business model and economics, not just sector label. Screen on: revenue model, end markets, size within an order of magnitude, growth profile, geographic mix, and capital intensity.

Search to validate: "[Company] competitors peer group comparison", "[Company] closest comparable companies".

**Write down why each peer is in — and name the obvious candidates you excluded, with the reason.** A comps table is only as good as its peer set, and an unstated peer set is unfalsifiable.

### Step 3 — Pull peer data

Run `find_securities` then `bigdata_company_tearsheet` for each peer. Pull the **same** metrics for everyone, from the same period, so the table is like-for-like. Note any fiscal-year misalignment.

### Step 4 — Build the comparables table

| Category | Metrics |
|----------|---------|
| Valuation | EV/Sales, EV/EBITDA, P/E (NTM and TTM), FCF yield, plus the sector-standard multiple |
| Growth | Revenue growth (TTM, NTM consensus), EPS growth |
| Profitability | Gross margin, EBITDA margin, operating margin, FCF margin |
| Returns | ROIC, ROE |
| Leverage | Net debt/EBITDA, interest coverage |
| Sentiment | Mean price target vs spot, rating distribution, quantified sentiment where available |

Use the multiples that fit the business — P/TBV for banks, P/AFFO for REITs, EV/Sales for pre-profit growth. Framework: [references/multiples-framework.md](./references/multiples-framework.md). Sector-specific KPIs: [references/sector-routing.md](./references/sector-routing.md).

### Step 5 — Position against the set

For each metric: the subject's value, the **peer median**, and its **quartile**. Percentile positioning shows what a raw table hides.

### Step 6 — Decompose the premium or discount

The core analytical step. If the company trades at a premium or discount, ask **what fundamentals justify it** — faster growth, higher margins, better returns, lower leverage, cleaner accounting — and how much of the gap remains **unexplained**. An unexplained gap is where the opportunity or the warning sits.

### Step 7 — Verdict

State relative attractiveness with the specific drivers, plus what would close or widen the gap.

Run [scripts/peer_comparables.py](./scripts/peer_comparables.py) only when the user explicitly wants a scripted comp table.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or spreadsheet-style version if useful.

## Quality bar

Non-negotiables:

- Peer set **justified**, with exclusions named — this is the credibility of the whole deliverable
- Same metrics, same period, for every company; fiscal misalignment flagged
- Multiples chosen for the business type, not generic P/E across a mixed set
- Peer median **and** quartile positioning given, not just raw values
- Premium/discount **decomposed** into justified and unexplained
- A relative verdict stated, with what would close the gap

Referenced files: 7

View saved version →

---
name: bigdata-post-ipo-day1
description: >
  Write a first-trading-day post-IPO reaction note for a newly listed company using Bigdata.com
  data plus web market data. Anchors the deal (offer price vs range, shares, greenshoe, implied
  market cap), reconstructs day 1 (open, intraday range, close, volume, first-day return),
  reads demand and float mechanics including stabilization, resets valuation against peers at
  the close, notes the quiet-period coverage gap, and maps the dated post-IPO timeline.
  Balanced, no buy/avoid call. Triggers: "post-IPO day 1", "first day trading reaction for X",
  "how did X's IPO debut", "X IPO pop", "X first day of trading", "IPO debut analysis".
---

# Bigdata Post-IPO Day 1 — First-Trading-Day Reaction

Assess the **first trading day** of a newly listed company: how price discovery played out against the offer, what demand and stabilization signals say, and how the stock is set up for the weeks ahead. Use Bigdata.com plugin tools plus web search for market data.

Read [references/post-ipo-common.md](./references/post-ipo-common.md) first — scope rules, data foundation, reference math, and the verify checklist apply in full.

**Use this skill when** the company has listed and is on or near day 1. Not this skill when:

| Request | Use instead |
|---------|-------------|
| The company has not yet priced | Pre-IPO analysis |
| ~Day 14, index inclusion question | Post-IPO day 14 |
| ~Day 179, 180-day lock-up expiry | Post-IPO day 179 |
| ~Day 365, founder lock-up and float expansion | Post-IPO day 365 |

Confirm the listing date and compute the trading-day count before committing to this workflow.

## What this note answers

- Where did the deal price (above / within / below range) and how did day 1 trade against it?
- Is the first-day move demand-driven, stabilization-supported, or thin-float mechanics?
- What do the open, close, and intraday range imply about the new valuation vs peers?
- What are the dated catalysts that now define the post-IPO timeline?

## Workflow

### Step 1 — Anchor the deal

From the prospectus / 424B and the pricing press release: final offer price, the range, shares offered (primary vs secondary), greenshoe size, total raised, implied market cap and EV at the offer, underwriters, listing date, ticker, exchange.

### Step 2 — Reconstruct day 1 (web search for market data)

Opening print, intraday high and low, first-day close, total volume and turnover vs shares offered, and the **first-day return** = (close − offer) / offer. Note any disclosed underwriter **stabilization / greenshoe** activity and whether the stock held above the offer.

### Step 3 — Demand and float mechanics

Free float as a % of shares outstanding — a small float amplifies moves. Retail vs institutional demand signals, oversubscription commentary, cornerstone and anchor behavior. Flag explicitly if the move is more about **scarce float** than fundamental demand.

### Step 4 — Valuation reset at the close

Recompute EV/Sales (and EV/EBITDA or P/E if profitable) at the first-day close, versus the offer and versus 3–6 listed peers. State where it screens rich or cheap — **without** a fair-value target.

### Step 5 — Sentiment and coverage

Bigdata.com sentiment and media reaction over the first day(s). Underwriter analysts are still in the **quiet period**, so there are no sell-side ratings yet — say so rather than implying a consensus that doesn't exist.

### Step 6 — Map the post-IPO timeline

Lay out the dated watch points: quiet-period end and first analyst initiations (~day 25), potential NASDAQ-100 fast-track inclusion (~day 15), first earnings report, and the 180-day and 366-day lock-up expiries.

## Output

Follow [assets/report-template.md](./assets/report-template.md).

- Length: 4–7 pages — this is a single-catalyst note.
- Cover line: company name, "Post-IPO — First Trading Day", date, "Prepared with Claude".
- Inline citations `[1]`, `[2]` after every sourced claim, hyperlinked to the document URL. Brand Bigdata.com content exactly "Bigdata.com".
- Full **Sources** section, then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Suggested filename: `Post_IPO_Day1_<Company>_<YYYY-MM-DD>` in the user's requested format.

## Quality bar

Run the verify checklist in [references/post-ipo-common.md](./references/post-ipo-common.md) before delivering. In particular:

- Every number traces to a recorded source or a labeled, shown calculation
- Float % + locked % reconcile to shares outstanding; market cap = price × shares outstanding
- The trading-day count matches the listing date
- No recommendation language ("we recommend", "attractive entry", "avoid", "buy the dip")
- Balanced bull/bear read and dated watch points only — no price target, no conviction rating

Referenced files: 5

View saved version →

---
name: bigdata-post-ipo-day14
description: >
  Write a day-14 post-IPO note on potential NASDAQ-100 fast-track index inclusion for a recently
  listed large-cap, using Bigdata.com data plus web search for index methodology and market data.
  Covers two-week trading status, an eligibility check against Nasdaq's current published rules
  (cited, never assumed), a float-adjusted index weight and implied passive-demand estimate with
  the math shown, days-to-cover versus ADV, the historical index effect and reversal risk, and
  dated watch points. Balanced, no buy/avoid call. Triggers: "NASDAQ-100 fast track for X",
  "index inclusion impact on X", "post-IPO day 14", "will X be added to the Nasdaq-100",
  "passive flows from index inclusion".
---

# Bigdata Post-IPO Day 14 — NASDAQ-100 Fast-Track Inclusion

A major IPO can qualify for **fast-track inclusion** in the NASDAQ-100 after a short minimum trading period, with inclusion effective around day 15. Run this note around **day 14** to capture the stock's status and the potential impact **before** inclusion takes effect. Use Bigdata.com plugin tools plus web search.

Read [references/post-ipo-common.md](./references/post-ipo-common.md) first — scope rules, data foundation, reference math, and the verify checklist apply in full.

> **Verify the rule, don't assume it.** NASDAQ-100 fast-entry eligibility (minimum trading days, the market-cap-rank threshold — historically top ~25% of the index) and the effective date are set by Nasdaq's published index methodology and can change. Confirm the current criteria and the specific effective date via web search before drawing conclusions, and state them as cited facts.

**Use this skill when** the company listed ~2 weeks ago and index inclusion is the live question. Not this skill when:

| Request | Use instead |
|---------|-------------|
| The company has not yet priced | Pre-IPO analysis |
| Day-1 debut reaction | Post-IPO day 1 |
| ~Day 179, 180-day lock-up expiry | Post-IPO day 179 |
| ~Day 365, founder lock-up and float expansion | Post-IPO day 365 |

## What this note answers

- Does the stock plausibly meet the fast-track eligibility criteria (market-cap rank, liquidity, seasoning)?
- If included, how large is the likely index weight and the mechanical passive demand?
- How does that demand compare to average daily volume — the "index effect" magnitude?
- How is the stock trading two weeks in, and what is already priced in?

## Workflow

### Step 1 — Two-week trading status

Price vs the offer and vs the day-1 close, trend and volatility, average daily volume (ADV), and current free float. Summarize whether the deal is working or fading.

### Step 2 — Eligibility check (cite the methodology)

From Nasdaq's **current** index methodology: minimum trading period, market-cap threshold, and liquidity/float requirements. Compare the company's market cap to the smallest current NASDAQ-100 constituents to gauge where it would rank. State eligibility as **likely / borderline / unlikely**, with the source.

### Step 3 — Passive-demand estimate (show the math, label as estimate)

- **Float-adjusted index weight** = company float-adjusted market cap / total NASDAQ-100 float-adjusted market cap
- **Implied passive buying** ≈ index weight × AUM tracking the NASDAQ-100 (QQQ plus other trackers — cite the AUM figure) / price = shares passive funds must buy
- **Days-to-cover** = implied passive buying / ADV — the core index-effect magnitude

### Step 4 — The index effect

Additions often drift up into the effective date as funds and front-runners accumulate, sometimes partially reversing afterwards. Cite recent NASDAQ-100 fast-track additions as analogs and how they traded around the event.

### Step 5 — Sentiment and risks

Bigdata.com sentiment and positioning. Risks both ways: inclusion not granted or delayed, the move already priced in, post-event reversal, or a high float/ADV diluting the effect.

### Step 6 — Watch points

Effective inclusion and rebalance date, rebalance mechanics, the 180- and 366-day lock-up expiries, and first earnings.

## Output

Follow [assets/report-template.md](./assets/report-template.md).

- Length: 4–7 pages — this is a single-catalyst note.
- Cover line: company name, "Post-IPO — NASDAQ-100 Fast-Track Inclusion", date, "Prepared with Claude".
- Inline citations `[1]`, `[2]` after every sourced claim, hyperlinked to the document URL. Brand Bigdata.com content exactly "Bigdata.com".
- Full **Sources** section, then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Suggested filename: `Post_IPO_Day14_IndexInclusion_<Company>_<YYYY-MM-DD>` in the user's requested format.

## Quality bar

Run the verify checklist in [references/post-ipo-common.md](./references/post-ipo-common.md) before delivering. In particular:

- Eligibility rules cited to Nasdaq's current methodology, never asserted from memory
- Every estimate labeled as an estimate, with its inputs and arithmetic shown
- AUM figure sourced, not assumed
- Days-to-cover computed against a stated ADV
- No recommendation language, no price target, no conviction rating

Referenced files: 5

View saved version →

---
name: bigdata-post-ipo-day179
description: >
  Write a day-179 post-IPO note on the 180-day lock-up expiry using Bigdata.com data plus
  filings and market data. Covers lock-up terms from the prospectus (expiry date, covered
  holders, share count, early-release provisions), float and overhang math (post-expiry float,
  days-to-trade versus ADV), insider and VC selling-intention signals, positioning into the
  event (short interest, borrow, options skew), the historical lock-up-expiry effect with
  analogs, and a two-sided read. Balanced, no buy/avoid call. Triggers: "180-day lock-up expiry
  for X", "lockup expiration impact", "post-IPO day 179", "shares unlocking for X",
  "insider selling after lockup", "float expansion at lockup".
---

# Bigdata Post-IPO Day 179 — 180-Day Lock-Up Expiry

The standard **180-day lock-up** is about to release insider, employee, and pre-IPO investor shares. Run this note around **day 179** to size the potential supply overhang and how the stock is set up into the unlock. Use Bigdata.com plugin tools plus web search.

Read [references/post-ipo-common.md](./references/post-ipo-common.md) first — scope rules, data foundation, reference math, and the verify checklist apply in full.

> **Confirm the terms from the filing.** Lock-up duration, exact expiry date, covered holders, the share count released, and any **early-release provisions** (price or time triggers, underwriter waivers) are issuer-specific. Pull them from the prospectus and any subsequent 8-K/424B — do not assume a generic 180 days for every holder.

**Use this skill when** the first lock-up tranche is about to expire. Not this skill when:

| Request | Use instead |
|---------|-------------|
| The company has not yet priced | Pre-IPO analysis |
| Day-1 debut reaction | Post-IPO day 1 |
| ~Day 14, index inclusion question | Post-IPO day 14 |
| The later 366-day founder tranche | Post-IPO day 365 |

## What this note answers

- How many shares unlock, and how much does free float expand?
- How large is the overhang relative to the stock's capacity to absorb it (days-to-trade)?
- Are insiders and VCs signaling intent to sell (secondary filings, 10b5-1 plans)?
- How has positioning — short interest, borrow, options skew — set up into the date?

## Workflow

### Step 1 — Lock-up terms recap (cite the filing)

Expiry date, who is locked (founders, employees, pre-IPO VC/PE, strategic holders), the **share count releasing**, and any tiered or early-release provisions. Distinguish this 180-day tranche from a later founder tranche if the deal staggered its lock-ups.

### Step 2 — Float and overhang math (show inputs)

- Current free float vs **post-expiry float** = current float + newly unlocked shares. Express the increase as a multiple of current float and as a % of shares outstanding.
- **Days-to-trade** = newly unlocked shares / ADV — the headline overhang measure.

### Step 3 — Selling-intention signals

Search for disclosed secondary offerings, 10b5-1 trading plans, prior insider sales, and management or VC commentary. VC funds near end-of-life are more likely distributors than long-horizon founders.

### Step 4 — Positioning into the event

Short interest and days-to-cover, borrow availability and cost, options skew, and recent price action. Heavy pre-positioning can mean the overhang is already partly discounted.

### Step 5 — Historical lock-up effect

Cite the typical pattern — often a modest negative drift into and just after expiry, scaled by the float increase, insider ownership, and post-IPO performance — plus 1–2 recent analogs.

### Step 6 — Two-sided read and watch points

**Bear:** dilution of tradable supply, insider distribution. **Bull:** overhang already priced, expiry as a clearing event, float increase aiding index eligibility and liquidity. Watch points: exact expiry date, any early release or waiver, follow-on secondary, next earnings.

## Output

Follow [assets/report-template.md](./assets/report-template.md).

- Length: 4–7 pages — this is a single-catalyst note.
- Cover line: company name, "Post-IPO — 180-Day Lock-Up Expiry", date, "Prepared with Claude".
- Inline citations `[1]`, `[2]` after every sourced claim, hyperlinked to the document URL. Brand Bigdata.com content exactly "Bigdata.com".
- Full **Sources** section, then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Suggested filename: `Post_IPO_Day179_LockUp_<Company>_<YYYY-MM-DD>` in the user's requested format.

## Quality bar

Run the verify checklist in [references/post-ipo-common.md](./references/post-ipo-common.md) before delivering. In particular:

- Lock-up terms cited to the filing, including early-release provisions — never assumed
- Float and overhang math shown with its inputs; float % + locked % reconcile to shares outstanding
- Days-to-trade computed against a stated ADV
- Both sides framed — the bull read on a clearing event is not optional
- No recommendation language, no price target, no conviction rating

Referenced files: 5

View saved version →

---
name: bigdata-post-ipo-day365
description: >
  Write a day-365 post-IPO note on the 366-day founder and significant-investor lock-up expiry
  and float expansion toward 15-20%, using Bigdata.com data plus filings and market data. Covers
  the staggered lock-up structure from the prospectus, float expansion math and days-to-trade,
  the offsetting float-adjusted index reweight demand netted against new supply, a realistic
  read on whether founders actually sell, the dual-class governance angle, and a two-sided
  setup. Balanced, no buy/avoid call. Triggers: "366-day lock-up for X", "founder lock-up
  expiry", "float expansion for X", "post-IPO one year lockup", "index reweight after float
  increase", "founder selling after IPO".
---

# Bigdata Post-IPO Day 365 — Founder Lock-Up & Float Expansion

A later **366-day lock-up** tranche releases founder and significant-investor shares, potentially expanding free float toward **15–20%**. Run this note around **day 365** to prepare for the supply increase and the offsetting mechanical demand. Use Bigdata.com plugin tools plus web search.

Read [references/post-ipo-common.md](./references/post-ipo-common.md) first — scope rules, data foundation, reference math, and the verify checklist apply in full.

> **Confirm the staggered structure from the filing.** Multi-tranche lock-ups, exactly which holders the 366-day tranche covers, the share count, and dual-class voting implications are issuer-specific. Pull them from the prospectus and subsequent filings — do not assume a standard structure.

**Use this skill when** the second, founder-level tranche is about to expire. Not this skill when:

| Request | Use instead |
|---------|-------------|
| The company has not yet priced | Pre-IPO analysis |
| Day-1 debut reaction | Post-IPO day 1 |
| ~Day 14, index inclusion question | Post-IPO day 14 |
| The earlier 180-day tranche | Post-IPO day 179 |

## What this note answers

- How far does free float expand (toward 15–20%?), and how many founder/investor shares become eligible?
- Does float-adjusted index re-weighting create mechanical buying that absorbs some of the new supply?
- How likely are founders and long-horizon investors to actually sell, versus diversify gradually?
- Does the founder retain control through super-voting shares even after selling economic stake?

## Workflow

### Step 1 — Lock-up structure recap (cite the filing)

The 366-day tranche: covered holders (founders, major VC/PE, strategic), the share count releasing, and how it stacks on the earlier 180-day unlock. Note any dual-class or super-voting structure.

### Step 2 — Float expansion math (show inputs)

- **Projected post-expiry float** = current float + newly eligible founder/investor shares. Express as a % of shares outstanding and check it against the 15–20% expectation.
- **Days-to-trade** = newly eligible shares / ADV. Treat this as a **ceiling** — founders rarely sell all at once.

### Step 3 — The offsetting demand: float-adjusted index reweighting

Many indices weight by **free-float** market cap, so a float increase raises the company's float-adjusted weight and triggers **mechanical passive buying** at the next reweight. Estimate the weight change and implied passive demand — show the math, label it an estimate — and **net it against the new supply**.

### Step 4 — Realistic supply assessment

Founders typically diversify gradually, often via 10b5-1 plans, rather than dump. End-of-life VC funds may distribute to LPs instead. Search for disclosed plans, prior selling behavior, and management commentary.

### Step 5 — Governance angle

Whether founders retain voting control via super-voting shares after selling economic stake, and what that means for the alignment narrative.

### Step 6 — Two-sided read and watch points

**Bear:** meaningful new founder and VC supply, plus the signaling effect. **Bull:** float increase improves liquidity, index weight, and institutional accessibility; passive demand offsets supply. Watch points: exact expiry date, index reweight date, disclosed sale plans, next earnings.

## Output

Follow [assets/report-template.md](./assets/report-template.md).

- Length: 4–7 pages — this is a single-catalyst note.
- Cover line: company name, "Post-IPO — Founder Lock-Up & Float Expansion", date, "Prepared with Claude".
- Inline citations `[1]`, `[2]` after every sourced claim, hyperlinked to the document URL. Brand Bigdata.com content exactly "Bigdata.com".
- Full **Sources** section, then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Suggested filename: `Post_IPO_Day365_FounderLockUp_<Company>_<YYYY-MM-DD>` in the user's requested format.

## Quality bar

Run the verify checklist in [references/post-ipo-common.md](./references/post-ipo-common.md) before delivering. In particular:

- Staggered lock-up structure cited to the filing — never assumed
- Supply and index-demand both quantified, with arithmetic shown, then **netted** against each other
- Days-to-trade presented as a ceiling, not a forecast
- Governance and voting control addressed explicitly where a dual-class structure exists
- No recommendation language, no price target, no conviction rating

Referenced files: 5

View saved version →

---
name: bigdata-pre-ipo-analysis
description: >
  Produce a balanced pre-IPO research note on an upcoming, not-yet-listed company using its
  S-1/F-1 plus Bigdata.com data. Covers deal structure (price range, shares, greenshoe, implied
  valuation, underwriters, lock-ups, share classes), two years plus interim financials, business
  model and funding history, TAM and listed comparables, IPO-window conditions, and 90-day
  sentiment — closing with bull and bear debates and watch points, never a participate/avoid
  call. Triggers: "analyze the IPO of X", "S-1 analysis", "upcoming listing for X", "IPO report
  on X", "should I look at X's IPO", "pre-IPO research on X", "X IPO valuation".
---

# Bigdata Pre-IPO Analysis

Institutional-style research note on an **upcoming** listing. Use Bigdata.com plugin tools plus web search for filings and market data.

**Use this skill when** the company has not yet priced. Not this skill when:

| Request | Use instead |
|---------|-------------|
| The company already listed and is trading | Post-IPO day 1 / 14 / 179 / 365 |
| An established public company's valuation | Valuation snapshot |
| A recommendation on whether to participate | Nothing — this deliverable is balanced by design |

## Scope rules (non-negotiable)

- **Upcoming IPOs only.** If the company has already listed, say this skill covers pre-listing analysis and offer a post-IPO note instead before proceeding.
- **Balanced framing only.** Never give a participate/wait/avoid recommendation, price target, or conviction rating. Present bull case, bear case, and watch points; let the reader decide.
- **No invented data.** If a figure (price range, offer size) is not yet public, write "not yet disclosed" rather than estimating. Label every third-party estimate as such.

## Data foundation (plugin tools + web)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_search` | Company background, IPO window conditions, sentiment | None |
| `find_securities` | Entity resolution when a tearsheet is needed | None |
| `bigdata_company_tearsheet` | Financial baseline where the entity is covered | `find_securities` |
| Web search | S-1/F-1 terms, financials, comparables, recent debuts | None |

**Fallback:** if Bigdata.com tools are unavailable, complete every step with web search alone and note in the footer that sentiment data was limited to public news.

## Workflow

### Step 1 — Clarify the input

Company name is required. If ambiguous, confirm with the user. Note the expected exchange and geography if known.

### Step 2 — Research (complete BEFORE building the report)

Run searches in this order, one focus and one time period per search. Record source name and date for every material fact as you go.

**a. Filing facts (web).** Latest S-1/F-1/prospectus: price range, shares offered (primary vs secondary), greenshoe, implied valuation, underwriters, expected pricing and listing date, exchange, ticker, use of proceeds, lock-up terms, share-class structure, cornerstone investors.

**b. Financials (web + filing).** Two most recent fiscal years plus the latest interim period: revenue, gross margin, operating income/loss, net income, operating cash flow, FCF, cash and debt.

**c. Company background (`bigdata_search` + web).** Business model, segments, customers, management, funding history and last private-round valuation.

**d. Industry and peers (web).** TAM estimates, competitive set, and 3–6 listed comparables with current EV/Sales, EV/EBITDA, or P/E as applicable.

**e. IPO window (`bigdata_search` + web).** Current IPO market conditions, recent debuts in the same sector, and how they traded in the aftermarket.

**f. Sentiment (Bigdata.com).** News flow and sentiment on the issuer over the last 90 days.

### Step 3 — Build the report

Follow [assets/report-template.md](./assets/report-template.md). Do not start document generation until research is complete.

### Step 4 — Verify before delivering

- Every number traces to a recorded source
- Internal consistency: implied valuation = price × post-offering shares outstanding
- All template sections present
- No recommendation language slipped in ("we recommend", "attractive entry", "avoid")

## Output

- Length: 6–10 pages.
- Cover: company name, "Pre-IPO Research Note", date, "Prepared with Claude".
- Add inline citations `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL. Brand Bigdata.com content exactly "Bigdata.com", linked to the `url` from the `bigdata_search` response.
- Full **Sources** section, then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown; offer PDF, Word (.docx), or presentation output.

## Quality bar

Non-negotiables in every pre-IPO note:

- Deal structure sourced from the filing, not from press summaries
- "Not yet disclosed" used wherever the filing is silent — never an estimate presented as fact
- Comparables named with their multiples, so the valuation framing is checkable
- Bull and bear both specific and falsifiable
- No participate/avoid call, no price target, no conviction rating
- Facts separated from analysis and implications

Referenced files: 4

View saved version →

---
name: bigdata-quick-take
description: >
  Give a fast, PM-style quick take on a stock using Bigdata.com data — a one-line current view,
  the 2-3 drivers that actually matter right now, the key risks and what would change the view,
  and the near-term setup with the next catalyst. Deliberately short: one page, no full thesis,
  no model. Triggers: "quick take on X", "what do you think of X", "give me a fast view on X",
  "thoughts on X", "X in a nutshell", "one-liner on X", "is X interesting right now".
---

# Bigdata Quick Take

One page, PM-style. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants a fast view, not a document. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Full thesis with recommendation and conviction | Investment memo |
| 30 days of developments, categorized | Company brief |
| "What is it worth" | Valuation snapshot |
| Risks rated by likelihood and impact | Risk assessment |
| Analysis around an earnings event | Earnings preview / digest |

**The discipline of this deliverable is brevity.** If the answer is running past a page, the user asked for a different skill.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Financial baseline, estimates, sentiment | `find_securities` |
| `bigdata_search` | What's live on the name right now | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Baseline

Call `bigdata_company_tearsheet` for the financial and valuation baseline plus sentiment. One pass, no exhaustive extraction.

### Step 3 — What's live (2–4 searches, not ten)

- "[Company] recent developments last 30 days"
- "[Company] analyst view valuation debate"
- "[Company] risks concerns"

Enough to know what the market is arguing about. Stop there.

### Step 4 — Filter to what matters

Keep the **2–3 drivers** that actually move the name now. Apply the EPIC lens quickly — is it material, can you form a view, does consensus miss it? — without writing the table out. Depth if needed: [references/epic-framework.md](./references/epic-framework.md).

### Step 5 — Take a view

State a current view in one line. Name the key risks and **what would change the view**. Give the near-term setup and the next catalyst with its date.

## Output

Follow [assets/report-template.md](./assets/report-template.md).

- Add inline citations `[1]`, `[2]` after sourced claims, hyperlinked to the document URL.
- End with **Sources**, then the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Markdown by default. A quick take rarely needs a Word document — offer only if the user asks.

## Quality bar

Pass the PM test: **What's different?** **What matters?** **What should I do about it?** (net assessment, key risk, next catalyst — no position sizing).

Non-negotiables:

- A **view** is actually stated — "it depends" is not a quick take
- 2–3 drivers maximum
- What would change the view is named and falsifiable
- Next catalyst has a date, or is flagged as undated
- One page. Brevity is the product.

Referenced files: 5

View saved version →

---
name: bigdata-regional-comparison
description: >
  Compare regions or blocs using Bigdata.com data — economic indicators, market performance,
  and cross-asset views — and turn that into an allocation recommendation. Covers growth,
  inflation, policy and labor per region, comparative developed-versus-emerging analysis,
  regional equity valuations, and fixed income and currency views for each. Triggers:
  "compare US vs Europe vs Asia", "which regions look attractive", "regional allocation",
  "developed vs emerging markets", "Europe vs US equities", "global allocation view".
---

# Bigdata Regional Comparison

Cross-region economic and market comparison with an allocation call. Use Bigdata.com plugin tools for every fact.

**Use this skill when** two or more regions or blocs are being weighed. Not this skill when:

| Request | Use instead |
|---------|-------------|
| One country in depth | Country analysis |
| The G7 specifically | G7 comparison |
| Sectors rather than regions | Cross-sector comparison |
| A sector inside one region | Country-sector analysis |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_country_tearsheet` | Economic data and comparisons where available | None |
| `bigdata_search` | Regional indicators, comparative analysis, cross-asset views | None |

If `bigdata_country_tearsheet` is unavailable or fails, complete the analysis with `bigdata_search` alone.

## Workflow

### Step 1 — Economic data for each region

Search each region separately, indicator by indicator.

**US:** "United States GDP growth economic outlook 2026" · "US inflation Federal Reserve interest rates" · "US unemployment labor market"

**Europe:** "Eurozone GDP growth economic outlook 2026" · "ECB interest rates inflation monetary policy" · "Europe unemployment economic data"

**Asia:** "Japan GDP growth BOJ monetary policy" · "China economic outlook GDP growth" · "India economic growth outlook"

Extend or substitute regions to match what the user asked for.

### Step 2 — Comparative analysis

- "G7 economic comparison GDP growth rates"
- "US Europe Asia economic outlook comparison"
- "developed vs emerging markets allocation"
- "global economic outlook regional comparison"

### Step 3 — Regional market implications

- "regional equity valuations US Europe Asia"
- "currency outlook major currencies USD EUR JPY"
- "global fixed income yields comparison"

### Step 4 — Cross-asset views

Build the fixed income and currency view for **each** region in scope, not just for equities. Regional allocation decisions are usually made across assets.

### Step 5 — Allocation call

Rank the regions and state the allocation explicitly, with the reason per region and the key risk that would break the call.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

Non-negotiables:

- Every region covered on the **same** indicators, so the comparison is like-for-like
- Policy divergence addressed explicitly — it usually drives currency and relative returns
- Cross-asset (equity, rates, FX) view per region, not equity-only
- An allocation call actually made, with the risk that would break it
- Every claim from a source carries an inline citation and appears in the Sources table

Referenced files: 4

View saved version →

---
name: bigdata-risk-assessment
description: >
  Produce a comprehensive risk assessment for a public company using Bigdata.com data (10-K risk
  factors, 8-K material events, news, tearsheet financials). Covers six categories — regulatory
  and legal, competitive and moat erosion, operational, financial and balance sheet, macro, and
  management and governance — each rated by likelihood and impact, with a distress screen when
  leverage is stretched, mitigation status, a priority matrix, and a scenario bridge to value
  drivers. Triggers: "risk assessment for X", "assess risks for X", "what are the risks with X",
  "what could go wrong at X", "risk factors for X", "how risky is X", "downside risks for X".
---

# Bigdata Risk Assessment

Comprehensive, evidence-rated risk profile of a public company. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants risks identified, rated, and prioritized. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Recent developments over the past month | Company brief |
| Near-term drivers around a specific print | Earnings preview / digest |
| Accounting and manipulation red flags specifically | Earnings quality screen |
| Moat durability and management quality specifically | Moat & governance review |
| Full thesis with a recommendation | Investment memo |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Leverage, liquidity, cash flow, coverage, debt maturity | `find_securities` |
| `bigdata_search` | 10-K risk factors, 8-K events, news, governance signals | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Before you synthesize — prioritize

Rate everything, but **lead with what moves the name**. A long list of low-likelihood, low-impact risks is a data dump. Identify the 2–3 risks that dominate the debate and put them first in the executive summary and the priority matrix.

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Financial health baseline

Call `bigdata_company_tearsheet` with the `entity_id` and analyze:

- **Leverage** — debt/equity, debt/assets, debt/EBITDA
- **Liquidity** — current ratio, quick ratio, cash position
- **Cash flow** — operating cash flow and free cash flow trends
- **Debt maturity** — short-term vs long-term
- **Interest coverage** — ability to service debt

### Step 3 — Distress quick screen (conditional)

When leverage is elevated, coverage is thin, FCF is weak against debt service, or liquidity is tight, add a quantitative flag:

- **Altman Z-Score (simplified)** — compute from tearsheet or filing inputs (working capital, retained earnings, EBIT, equity, total liabilities, sales, total assets). If inputs are incomplete, state the **data gaps** and give a qualitative distress read instead.
- **Manipulation context** — if accruals or earnings quality look aggressive, work through [references/red-flags-checklist.md](./references/red-flags-checklist.md). Run [scripts/earnings_quality.py](./scripts/earnings_quality.py) only if the user explicitly wants scripted metrics.

### Step 4 — Moat and competitive durability

Identify the **moat type** (if any) and how it could erode, using [references/moat-taxonomy.md](./references/moat-taxonomy.md):

- Pricing power trend vs peers
- Share shifts to entrants or substitutes
- ROIC compression vs history and vs cost of capital

Search if needed: "[Company] pricing power market share competition ROIC".

### Step 5 — Official risk disclosures

Use `bigdata_search` for the company's own disclosures — the authoritative baseline:

- "risk factors material risks regulatory competitive in the last 10-K SEC filing of [Company]"

If the 10-K can't be found:

> "I couldn't locate the most recent 10-K filing. Should I proceed with 8-K filings and news-based risk analysis?"

### Step 6 — Material events (8-K)

- "material events changes risks in 8-K SEC filing of [Company] in the last 90 days"

8-Ks catch emerging risks that post-date the annual report.

### Step 7 — Emerging risks in the news

Run **4–6 targeted searches** across risk categories:

- "[Company] regulatory investigation lawsuit controversy in the last 30 days"
- "[Company] competitive pressure market share losses"
- "[Company] supply chain disruption operational challenges"
- "[Company] executive departure management changes"
- "[Company] cybersecurity breach data incident"

### Step 8 — Management and governance signals

Especially for founder-led or concentrated-ownership names:

- "[Company] CEO chairman combined role board independence"
- "[Company] insider selling stock compensation"
- "[Company] activist shareholder governance"
- "[Company] related party transactions"

Cross-check patterns against [references/capital-allocation.md](./references/capital-allocation.md).

### Step 9 — Categorize and rate

Sort every risk into the six categories and rate each on **likelihood** and **impact**:

**Likelihood** — High: >50% within 12 months or already materializing · Medium: 20–50% within 12–24 months · Low: <20% or >24 months out

**Impact** — High: >10% of revenue/earnings or existential · Medium: 3–10% or significant operational impairment · Low: <3% or manageable

**Combined priority** — High×High = Critical · High×Medium or Medium×High = High · Medium×Medium = Medium · otherwise Lower

The six categories:

1. **Regulatory/legal** — litigation, investigations, antitrust, framework changes, product liability
2. **Competitive** — moat erosion first: which moat, how it breaks; then share loss, entrants, pricing pressure (often the first signal of moat damage), customer concentration, ROIC vs WACC compression
3. **Operational** — supply chain, key personnel, technology and cyber, production constraints, execution
4. **Financial/balance sheet** — refinancing and covenants, liquidity, FX and rates, pensions, off-balance-sheet, plus the Step 3 distress read
5. **Macro/market** — cycle sensitivity, geopolitics, secular decline, commodities, policy
6. **Management & governance** — board independence, dual CEO/chair, related-party exposure, compensation design, insider patterns, capital allocation credibility, succession

Be **objective and evidence-based** in the ratings, note **mitigation status** for each material risk, and distinguish materiality — not every disclosed risk factor deserves equal weight.

### Step 10 — Scenario bridge

Connect the likelihood × impact view to a brief **bull / base / bear narrative** for the value drivers. Narrative, not a full DCF — this is what makes the assessment actionable rather than descriptive.

If the profile comes out thin:

> "Risk profile appears relatively low based on available information. Would you like me to expand the search parameters or focus on industry-specific risks?"

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, rating tables, sources, and footer.

- Add inline citations as superscript-style numbers `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version at the end if useful.

## Quality bar

Pass the PM test before delivering: **What's different?** **What matters (2–3 risks)?** **What should I do about it?** (net assessment, key risk, next catalyst — no position sizing).

Non-negotiables in every assessment:

- Official 10-K disclosures used as the baseline, then validated against 8-Ks and news
- Every material risk carries **likelihood, impact, evidence, and mitigation status**
- Moat erosion treated as a first-order competitive risk, not a footnote
- Distress screen run whenever leverage or coverage is weak
- Governance and management assessed explicitly, not folded into "operational"
- Scenario bridge present — descriptive risk lists are not actionable
- Facts separated from analysis and implications

Referenced files: 9

View saved version →

---
name: bigdata-scenario-analysis
description: >
  Build bull, base, and bear cases for a public company using Bigdata.com data — with explicit
  line-item assumptions, justified probability weights summing to 100%, a value or price per
  scenario with the bridge shown, a probability-weighted expected value and expected return
  versus spot, the upside/downside skew and risk-reward ratio, and what would move probability
  between the cases. Triggers: "scenario analysis for X", "bull base bear for X", "what's the
  upside and downside on X", "expected value for X", "probability-weighted view on X",
  "risk reward on X", "model out the cases for X".
---

# Bigdata Scenario Analysis

Three cases, honest probabilities, and the arithmetic. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants outcomes weighted, not a single point estimate. Not this skill when:

| Request | Use instead |
|---------|-------------|
| A single valuation read | Valuation snapshot |
| Full thesis with recommendation | Investment memo |
| Risks rated by likelihood and impact, not valued | Risk assessment |
| Scenarios specifically around a print | Earnings preview |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Financials, consensus estimates, multiples, spot price | `find_securities` |
| `bigdata_search` | The live debate, bull and bear arguments, analyst ranges | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company and baseline

Resolve the entity and pull the tearsheet: current financials, consensus estimates, current multiples, and **spot price** — every scenario is measured against it.

### Step 2 — Find the swing variables

Scenarios are only useful if they turn on the **2–3 variables that actually decide the outcome** — not on twenty inputs nudged in the same direction. Search the live debate:

- "[Company] bull case bear case debate"
- "[Company] key drivers revenue growth margin outlook"
- "[Company] analyst price target range high low"

Pick the swing variables and hold everything else roughly constant across cases. This is what makes the scenarios interpretable.

### Step 3 — Build the three cases

For each of **bull / base / bear**, state assumptions at the line-item level:

| Assumption | Bear | Base | Bull |
|------------|------|------|------|
| Revenue growth | | | |
| Operating margin | | | |
| [Swing variable 3] | | | |
| Exit multiple or terminal assumption | | | |

The **base case should be roughly consensus** — if it isn't, say so explicitly and explain why, because that gap is itself the finding.

### Step 4 — Value each scenario

Derive a value or price per case and **show the bridge** — the multiple applied to which earnings, or the DCF assumptions changed. Methodology: [references/dcf-methodology.md](./references/dcf-methodology.md), [references/reverse-dcf.md](./references/reverse-dcf.md).

### Step 5 — Assign and justify probabilities

Weights must sum to ~100%, and each needs a **one-line justification** grounded in evidence. Guard against the usual failure: a comfortable 25/50/25 that was never really thought about. If the distribution is skewed, say so. Methodology: [references/thesis-construction.md](./references/thesis-construction.md).

### Step 6 — Expected value and skew

- **EV** = Σ (probability × value). Show the arithmetic.
- **Expected return** versus spot, in %.
- **Upside/downside ratio** = (bull − spot) / (spot − bear).
- Note whether the distribution is symmetric or skewed, and what that means for the setup.

Run [scripts/scenario_probability.py](./scripts/scenario_probability.py) or [scripts/dcf_model.py](./scripts/dcf_model.py) only when the user explicitly asks for scripted math.

### Step 7 — What moves probability

For each case, name the **specific, observable** developments that would raise or lower its weight. Scenarios without triggers are static and go stale within a quarter.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) version if useful.

## Quality bar

Non-negotiables:

- 2–3 **swing variables** identified; everything else held roughly constant
- Assumptions at line-item level per case, not narrative adjectives
- Probabilities sum to ~100% and each is **justified**, not defaulted
- Base case tied to consensus, or the divergence stated explicitly
- Value bridge shown per scenario — no unexplained price targets
- EV arithmetic written out, plus expected return versus spot and the skew
- Probability triggers named and observable

Referenced files: 9

View saved version →

---
name: bigdata-sector-analysis
description: >
  Analyze a market sector using Bigdata.com data — performance, valuations, themes,
  sub-industries, and upcoming catalysts. Maps the sector to its own operating and valuation
  KPIs rather than generic P/E, reads cycle and profitability positioning as early, mid, or late
  versus history, aggregates bellwether tearsheet metrics, and closes with a positioning call
  plus top picks and areas to avoid. Triggers: "analyze the X sector", "what's happening in X
  sector", "X sector outlook", "how is the X industry doing", "X sector valuations",
  "is the X sector attractive", "semiconductor/energy/healthcare sector view".
---

# Bigdata Sector Analysis

Full read on one sector: where it trades, what drives it, and what is coming. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the subject is one sector. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Two or more sectors compared, or rotation | Cross-sector comparison |
| A sector inside a specific country or region | Country-sector analysis |
| An actionable KPI-and-debates playbook for investing the sector | Sector playbook |
| A macro theme that cuts across sectors | Thematic research |
| One company in the sector | Company brief / investment memo |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_search` | Sector trends, valuations, policy, catalysts, cycle context | None |
| `find_securities` | Entity ids for 5–10 sector bellwethers | None |
| `bigdata_company_tearsheet` | Per-company metrics, estimates, sentiment, segments | `find_securities` |
| `bigdata_events_calendar` | Upcoming earnings and conferences | `find_securities` |

Run **5–10 targeted searches** across the workflow. Include temporal context ("last 30 days", "2026 outlook").

## Workflow

### Step 1 — Sector context

Search:

- "[Sector] sector outlook trends analysis"
- "[Sector] sector earnings performance"
- "[Sector] sector headwinds tailwinds"
- "[Sector] sector valuations multiples"
- "[Sector] sector regulatory policy"

### Step 2 — Sector-specific KPI lens (GICS)

Do **not** rely only on generic P/E, P/S, and EV/EBITDA. Map the sector to its primary operating and valuation KPIs:

| GICS sector | Emphasize these KPIs |
|-------------|----------------------|
| Information Technology / Software-SaaS | ARR growth, NRR, Rule of 40, FCF margin, payback |
| Financials | NIM, CET1 / capital, credit costs, ROTCE, efficiency |
| Health Care (incl. Pharma) | Growth drivers, pipeline / patent, R&D, payer mix, regulatory |
| Real Estate (REITs) | AFFO, NAV, cap rates vs bonds, same-store NOI |
| Industrials | Backlog, book-to-bill, margin mix, OEM / capex cycle |
| Consumer Discretionary / Staples | Same-store sales, promo, input costs, private label |
| Energy | Commodity linkage, breakeven, FCF at forward curve, capital discipline |
| Materials | Price/volume, capacity, inventory, China / construction linkage |
| Communication Services | Subscribers, ARPU, churn, ad market / streaming economics |
| Utilities | Allowed ROE, rate case risk, weather / load growth |
| (Other) | Default to margin trajectory, ROIC vs peers, and segment growth |

Deeper playbooks: [references/sector-routing.md](./references/sector-routing.md).

### Step 3 — Key companies

Use `find_securities` for 5–10 major sector companies, then `bigdata_company_tearsheet` for each: financial metrics and performance, analyst estimates and sentiment, revenue segmentation, ESG scores.

### Step 4 — Aggregate sector metrics

From the tearsheets, compile sector-relevant multiples (per Step 2, not only P/E), the Step 2 KPIs where visible, revenue and earnings growth trends, the analyst rating distribution, and sentiment indicators.

### Step 5 — Cycle and profitability positioning

Add brief, evidence-based cycle context:

- Search "[Sector] sector ROIC profitability cycle outlook" and "[Sector] margin cycle vs history"
- State whether ROIC (or a sector proxy) and margins look **early / mid / late** versus a normal cycle — or flag the data limits
- Industry-economics mental model: [references/porter-five-forces.md](./references/porter-five-forces.md)

### Step 6 — Catalysts

Search:

- "[Sector] regulatory changes policy"
- "[Sector] technology disruption"
- "[Sector] M&A consolidation"
- "[Sector] earnings expectations"
- "[Sector] supply chain tariffs"

### Step 7 — Events calendar

Use `bigdata_events_calendar` for upcoming earnings and conferences across the bellwethers.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

Non-negotiables in every sector analysis:

- Sector-specific KPIs present — a report built only on P/E has not done the job
- Cycle positioning stated (early / mid / late) or its data limits flagged
- Tailwinds and headwinds name **which companies** are exposed
- Positioning call given: overweight / neutral / underweight, with top picks and areas to avoid
- Every claim from a source carries an inline citation and appears in the Sources table

## GICS sectors reference

Information Technology, Health Care, Financials, Consumer Discretionary, Consumer Staples, Industrials, Energy, Materials, Real Estate, Communication Services, Utilities.

Referenced files: 6

View saved version →

---
name: bigdata-sector-playbook
description: >
  Build an actionable investment playbook for a sector using Bigdata.com data and sector-specific
  frameworks — the KPIs that actually matter in that sector, how to value companies in it and
  why, the live debates and where consensus sits on each, valuation context against the sector's
  own history, a sub-industry map with cycle position, screening criteria and sector-specific
  red flags, and an actionable setup of what to own, avoid, and watch. More operational than a
  sector analysis: it teaches how to invest the sector, not just how it is doing. Triggers:
  "sector playbook for X", "how do I analyze X companies", "what KPIs matter in X",
  "how to value X sector companies", "investing framework for X sector", "X sector cheat sheet".
---

# Bigdata Sector Playbook

The operating manual for a sector: what to measure, what is debated, what to own. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants the framework for investing a sector. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Current performance, valuations, and catalysts | Sector analysis |
| Sectors ranked against each other | Cross-sector comparison |
| A sector inside one country | Country-sector analysis |
| One company against its peers | Peer comparables |

**Analysis vs playbook:** the sector analysis says how the sector is doing right now; the playbook says how to analyze any company in it, and what setup is actionable today.

## Sector references

Load the matching playbook for the sector in scope — routing table: [references/sector-routing.md](./references/sector-routing.md).

| Sector (GICS) | Reference |
|---------------|-----------|
| Information Technology (SaaS / software) | [references/technology-saas.md](./references/technology-saas.md) |
| Financials (banks, insurance) | [references/financials-banks.md](./references/financials-banks.md) |
| Health Care (pharma, biotech, devices) | [references/healthcare-pharma.md](./references/healthcare-pharma.md) |
| Real Estate (REITs) | [references/reits.md](./references/reits.md) |
| Industrials (A&D, machinery, transport) | [references/industrials.md](./references/industrials.md) |
| Consumer Discretionary / Staples | [references/consumer-retail.md](./references/consumer-retail.md) |
| Energy | [references/energy.md](./references/energy.md) |

If the fit is unclear: [references/sector-selection-guide.md](./references/sector-selection-guide.md). For sectors without a dedicated file (Materials, Communication Services, Utilities), build the KPI framework from the closest analog plus first principles — margin structure, capital intensity, cycle drivers — and say that you did.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_search` | Sector debates, valuation context, structural trends | None |
| `find_securities` | Entity ids for sector constituents | None |
| `bigdata_company_tearsheet` | KPI availability and current levels across the sector | `find_securities` |
| `bigdata_events_calendar` | Sector catalyst timing | `find_securities` |

## Workflow

### Step 1 — Load the sector reference

Read the matching file above **before** searching. It defines the KPI vocabulary the rest of the playbook uses.

### Step 2 — KPI framework

Set out the operating and valuation KPIs that matter in this sector, what "good" looks like for each, and where each is found. This is the core of the playbook — a reader should be able to pick up any company in the sector and know what to measure.

### Step 3 — Valuation approach

State how companies in this sector are valued and **why** that method fits the economics — P/AFFO and NAV for REITs, P/TBV and ROTCE for banks, EV/Sales with Rule of 40 for SaaS. Note where the standard method breaks down.

### Step 4 — The live debates

Search for what the sector is actually arguing about:

- "[Sector] investment debate bulls bears"
- "[Sector] structural change disruption outlook"
- "[Sector] margin sustainability capacity"

For each debate: state both sides, where consensus currently sits, and what evidence would settle it. Sector calls are usually made or lost on these.

### Step 5 — Valuation context

Where the sector trades versus its **own** history — not just versus the market. Multiples now against 5- and 10-year ranges, and on what earnings base (peak, mid-cycle, trough). Industry structure: [references/porter-five-forces.md](./references/porter-five-forces.md).

### Step 6 — Sub-industry map

Break the sector into sub-industries, and for each: economics, cycle position, and current setup. Sector-level averages usually hide the dispersion that matters.

### Step 7 — Screening criteria and red flags

- **Screen for:** the metrics that identify quality and value in this sector specifically
- **Red flags:** the sector's characteristic failure modes — channel stuffing in consumer, reserve releases in insurance, capitalized development in software, decline-rate masking in energy

### Step 8 — Actionable setup

Close with what to own, what to avoid, and what to watch — each tied to the KPIs and debates above, with named companies where the evidence supports it.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version if useful.

## Quality bar

Non-negotiables:

- KPI framework is **sector-specific** and usable on any company in the sector
- Valuation method justified by the economics, with its breaking point named
- Debates presented two-sided, with where consensus sits and what would settle them
- Valuation context against the sector's **own** history, on a stated earnings base
- Sub-industry dispersion shown, not averaged away
- Sector-characteristic red flags named
- An actionable setup delivered — own / avoid / watch

Referenced files: 14

View saved version →

---
name: bigdata-thematic-research
description: >
  Research a macro investment theme using Bigdata.com data — scope and sub-themes, investment
  implications, sector impact, named beneficiaries and vulnerable losers with tearsheet
  fundamentals, the policy and regulatory dimension, geographic impact, and concrete
  implementation ideas. Covers themes such as AI, energy transition, inflation and rates,
  deglobalization and reshoring, demographics, geopolitical risk, and fiscal policy. Triggers:
  "research the X theme", "X investment implications", "who benefits from X", "how do I play
  X", "AI / energy transition / deglobalization theme", "thematic view on X".
---

# Bigdata Thematic Research

Cross-sector research on one macro theme, ending in implementable ideas. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the subject is a theme that cuts across sectors or borders. Not this skill when:

| Request | Use instead |
|---------|-------------|
| One sector's performance and outlook | Sector analysis |
| Sectors ranked against each other | Cross-sector comparison |
| One country's economy | Country analysis |
| One company exposed to the theme | Company brief / investment memo |

Common themes: AI and technology transformation, energy transition and clean tech, inflation and interest rates, deglobalization and reshoring, demographic shifts, geopolitical risk, fiscal policy and government spending.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `bigdata_search` | Theme coverage, implications, policy, market impact | None |
| `find_securities` | Entity ids for the most exposed companies | None |
| `bigdata_company_tearsheet` | Fundamentals and exposure of beneficiaries and losers | `find_securities` |
| `bigdata_country_tearsheet` | Geographic impact where available | None |

## Workflow

### Step 1 — Define the theme scope

State the boundaries and the sub-themes explicitly before searching. An unbounded theme produces an unbounded report — this step is what keeps the deliverable usable.

### Step 2 — Search the theme (5–10 queries)

- "[Theme] investment implications outlook"
- "[Theme] winners beneficiaries stocks"
- "[Theme] risks losers vulnerable"
- "[Theme] policy government regulation"
- "[Theme] market impact analysis"
- "[Theme] sector exposure"

### Step 3 — Beneficiaries and casualties

Use `find_securities` and `bigdata_company_tearsheet` for the most exposed companies on **both** sides. A theme note that names only winners is a pitch, not research — quantify the exposure where the data allows (revenue share, capex tied to the theme, contract backlog).

### Step 4 — Geographic impact

Use `bigdata_country_tearsheet` (or search) for the countries most affected, positively and negatively.

### Step 5 — Implementation

Turn the analysis into concrete ways to express the theme: direct beneficiaries, second-order plays, avoided exposures, and what would invalidate the theme.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- **Inline citations** `[1]`, `[2]` after every claim from a source, hyperlinked to the document URL.
- End with the numbered **Sources** table (source, date, URL), then the **Powered by Bigdata.com** line and **Disclaimer**, verbatim.
- Default format is Markdown. After delivering, you may ask: "Would you like me to create a Word document or presentation with this analysis?"

## Quality bar

Non-negotiables:

- Theme scope and sub-themes stated up front and held to
- **Both** beneficiaries and losers named, with exposure quantified where possible
- Policy dimension addressed — most macro themes are policy-driven
- Implementation section present: how to express the theme, and what would invalidate it
- Every claim from a source carries an inline citation and appears in the Sources table

Referenced files: 4

View saved version →

---
name: bigdata-valuation-snapshot
description: >
  Answer what a public company is worth and whether it is cheap or expensive, using Bigdata.com
  data (tearsheet multiples, estimates, margins, peer context). Produces a multiples cross-check
  against the company's own history and peer median, an implied-expectations read on what the
  current price already embeds (reverse-DCF reasoning, no model build required), the 2-3 value
  drivers that dominate, and a cheap / fair / rich verdict. Triggers: "what is X worth", "is X
  expensive", "valuation snapshot for X", "what's priced in for X", "is X cheap vs peers",
  "how is X valued", "fair value for X".
---

# Bigdata Valuation Snapshot

The lightweight answer path for "what is it worth" — no full memo, no standalone model build. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the user wants a valuation read without a full thesis. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Full thesis with recommendation and conviction | Investment memo |
| Explicit bull/base/bear with probabilities and EV | Scenario analysis |
| Detailed peer table across many metrics | Peer comparables |
| Valuation in the context of an upcoming print | Earnings preview |
| Built DCF or sum-of-parts model output | Investment memo (with scripts) |

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Current and historical multiples, estimates, margins, FCF, segments | `find_securities` |
| `bigdata_search` | Peer valuation context, analyst views, valuation debates | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Identify the company

Call `find_securities` with the company name to get the `entity_id`.

### Step 2 — Pull valuation inputs

Call `bigdata_company_tearsheet` for current and historical multiples, consensus estimates, margins, FCF where shown, and segment context.

### Step 3 — Peer and history context

- Use the tearsheet peer set, or search: "[Company] valuation vs peers EV EBITDA PE comparison"
- Note **current** vs **~5-year range** or trailing average where the data allows. If only spot data exists, say so and approximate rather than inventing a range.
- Pick the multiples that fit the business — a bank on P/TBV, a REIT on P/AFFO, a pre-profit grower on EV/Revenue. Framework: [references/multiples-framework.md](./references/multiples-framework.md).

### Step 4 — Implied expectations (reverse-DCF mindset)

Without building a model, articulate **what has to go right** at the current price:

- Revenue growth the multiple embeds vs consensus
- Margin level or trajectory embedded vs recent trend
- Reinvestment needs and the risk premium implied
- Whether the market is pricing a re-rating or a de-rating vs fundamentals

Methodology: [references/reverse-dcf.md](./references/reverse-dcf.md). Full DCF mechanics if the user wants depth: [references/dcf-methodology.md](./references/dcf-methodology.md). Run [scripts/reverse_dcf.py](./scripts/reverse_dcf.py) or [scripts/dcf_model.py](./scripts/dcf_model.py) only when the user explicitly asks for scripted or spreadsheet-style output.

### Step 5 — Synthesize

Combine the **multiples cross-check**, the **implied expectations**, and the **2–3 value drivers** that actually move fair value. State plainly whether the stock screens **cheap, fair, or rich** relative to embedded expectations — and name what would change that.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations as superscript-style numbers `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) or presentation version at the end if useful.

## Quality bar

Pass the PM test before delivering: **What's different?** **What matters (2–3 drivers)?** **What should I do about it?** (net assessment, key risk, next catalyst — no position sizing).

Non-negotiables in every snapshot:

- Multiples chosen for the **business type**, not generic P/E on everything
- Current level always framed against **history and peers**, or the gap stated explicitly
- A plain-English statement of what the price embeds — this is the point of the deliverable
- Cheap / fair / rich verdict given, not hedged into nothing
- Tearsheet and search preferred over model builds; scripts only on request
- Facts separated from analysis and implications

Referenced files: 9

View saved version →

---
name: bigdata-variant-perception
description: >
  State explicitly where your view on a public company differs from consensus, using Bigdata.com
  data. Establishes the consensus baseline from estimates and sell-side posture, applies the
  EPIC filter to candidate differentiators, frames the view on FaVeS (fundamentals, valuation,
  sentiment), and states the variant view as a specific, falsifiable claim with a time horizon —
  plus what the market is missing, why the mispricing persists, the evidence for the view, and
  what would disprove it. Triggers: "variant perception on X", "where do I differ from
  consensus on X", "what is the market missing on X", "non-consensus view on X",
  "what's priced in versus reality for X", "contrarian case for X".
---

# Bigdata Variant Perception

The discipline of saying exactly where you differ from consensus — and how you'd know you were wrong. Use Bigdata.com plugin tools for every fact.

**Use this skill when** the consensus gap *is* the deliverable. Not this skill when:

| Request | Use instead |
|---------|-------------|
| Full thesis with recommendation and conviction | Investment memo |
| Absolute valuation | Valuation snapshot |
| Outcomes weighted by probability | Scenario analysis |
| A fast view with no consensus framing | Quick take |

**A variant perception is not a bull case.** Agreeing with consensus more enthusiastically is not a variant view. If you cannot name a specific number, timing, or outcome where you differ, the honest answer is that you have no variant perception on this name — say so.

## Data foundation (plugin tools)

| Tool | Purpose | Prerequisite |
|------|---------|--------------|
| `find_securities` | Resolve company name → RavenPack `entity_id` | None |
| `bigdata_company_tearsheet` | Consensus estimates, multiples, sentiment, positioning | `find_securities` |
| `bigdata_search` | Sell-side posture, the live debate, evidence for and against | None |

If the company name is ambiguous after `find_securities`, ask:

> "I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

## Workflow

### Step 1 — Establish the consensus baseline

You cannot differ from a consensus you have not written down. From the tearsheet and search, capture:

- Consensus revenue, EPS, and margin estimates for the next 1–2 years
- Mean price target and the high/low range
- Rating distribution and recent revision direction
- What the current multiple implies (reverse-DCF reasoning — [references/reverse-dcf.md](./references/reverse-dcf.md))

Search: "[Company] analyst estimates consensus outlook", "[Company] price target upgrades downgrades".

### Step 2 — Apply the EPIC filter

For each candidate differentiator, run all four tests:

| Test | Question | Pass criteria |
|------|----------|---------------|
| **E**ffect | Is it material? | ~10% change moves intrinsic value meaningfully |
| **P**redictability | Can you forecast it? | You have an analytical or informational edge, not a guess |
| **I**ndependence | Does consensus get it wrong? | The market systematically misjudges this |
| **C**onsensus gap | Is there a gap? | Your forecast differs meaningfully and specifically |

Only factors passing all four qualify. Detail: [references/epic-framework.md](./references/epic-framework.md).

### Step 3 — Frame on FaVeS

- **Fundamentals** — which 2–3 KPIs drive value, and where your forecast differs from the consensus line item
- **Valuation** — what multiple the quality and growth justify, versus what is being applied
- **Sentiment** — what is priced in behaviorally: positioning, flows, short interest, sell-side posture

Detail: [references/faves-framework.md](./references/faves-framework.md).

### Step 4 — State the variant view

Write it as a **specific, falsifiable claim with a time horizon**:

> "Consensus models [X]% [metric] in [period]; we expect [Y]% because [mechanism], which would imply [$Z] of [revenue/EBIT/value] versus the [$W] embedded in the current price."

Vague directional statements ("we're more optimistic than the street") fail this deliverable.

### Step 5 — Why the mispricing persists

A gap that anyone could see would already be closed. Name the structural reason it survives: disclosure gaps, time-horizon mismatch, index or mandate constraints, coverage gaps, complexity, recency bias after a shock, or a segment that reporting obscures. Methodology: [references/thesis-construction.md](./references/thesis-construction.md).

### Step 6 — Evidence and disconfirmation

- **Evidence for:** the specific data points, each cited
- **What would disprove it:** observable, dated, and specific — if nothing could disprove the view, it is not a research claim
- **Time horizon:** when the gap should close, and what closes it

Run [scripts/reverse_dcf.py](./scripts/reverse_dcf.py) only if the user explicitly wants scripted implied-growth math.

## Output

Follow [assets/report-template.md](./assets/report-template.md) exactly — section order, tables, sources, and footer.

- Add inline citations `[1]`, `[2]` immediately after claims, hyperlinked to the document URL.
- Every deliverable ends with the **Powered by Bigdata.com** line and the **Disclaimer**, verbatim.
- Default format is Markdown; offer a Word (.docx) version if useful.

## Quality bar

Non-negotiables:

- Consensus baseline **written down** with numbers before any differing view is stated
- EPIC run on each candidate; only all-four passes qualify
- The variant view is specific, quantified, and carries a time horizon
- A structural reason the mispricing persists — otherwise the gap probably isn't real
- Disconfirming evidence named and observable
- Honesty about the null result: if nothing passes EPIC, say there is no variant perception here

Referenced files: 9