Chronograph LP

View saved version →

---
name: chronograph-cashflow-forecast
description: Forecast private capital cashflows for existing portfolios using Chronograph MCP data and a Takahashi-Alexander style model. Use when Codex needs to analyze or forecast LP-level contributions, distributions, NAV, unfunded exposure, net cashflows, or Excel-style cashflow forecast outputs from existing Chronograph funds, commitments, groups, or portfolios.
---

# Chronograph Cashflow Forecast

**Requirements:** The Existing Portfolio Forecast mode requires a connected Chronograph MCP server — these workflows are designed for permissioned Chronograph users to connect to their private investment data. The Future Commitment Pacing Overlay mode runs from user-provided assumptions and works without a connection.

## Overview

Use the Chronograph MCP server as the source of truth for private capital commitment data, then forecast fund and portfolio cashflows with a Takahashi-Alexander style model. Prefer concise analysis in chat unless the user asks for a workbook, export, or model artifact.

Do not hard-code tool names — read the Chronograph MCP server's live tool descriptions to pick the appropriate tool for each data need.

## Workflow

1. Clarify scope only when needed: portfolio, group, fund IDs, commitment IDs, currency, as-of date, forecast horizon, and units. V1 is existing-portfolio-only; do not model future commitments or pacing schedules.
2. Resolve entities. The Chronograph MCP server exposes a tool for resolving fund, group, GP, or company names to IDs. Use it to turn user-provided names into IDs before pulling any metrics.
3. Pull fund metadata. The Chronograph MCP server exposes a generic query tool for retrieving core entity attributes — use it to get fund name, fund type, vintage year, reporting currency, geographic focus, general partner, and final reporting date where useful.
4. Pull net LP commitment values: NAV, called, distributed, unfunded, commitment amount, net IRR, net MOIC, DPI, and RVPI. The Chronograph MCP server exposes a dedicated tool for net LP-level performance that covers all of these. Treat the values as net and surface the as-of date, currency, and that the values are net (not gross) in user-facing results.
5. Map fund types to forecast assumptions. Use the default assumptions in `references/model-methodology.md` unless the user provides custom assumptions.
6. Build yearly or quarterly forecast periods from the as-of date through the requested horizon. Default to annual periods for executive analysis and quarterly periods for Excel-style output.
7. Forecast contributions, distributions, NAV, unfunded, and net cashflow by fund, then aggregate to portfolio, group, vintage, fund type, or GP as requested.
8. Present assumptions, source context, and checks. Always state currency, units, as-of date, forecast horizon, and that future commitments are excluded in V1.
9. If the user requests Excel, create a workbook with Inputs, Fund Forecast, Portfolio Summary, and Checks.

## Planning Mode

Use one of these modes, or combine them when the user asks for a liquidity plan:

- **Existing Portfolio Forecast.** Use Chronograph actuals for current commitments, NAV, unfunded, called, distributed, and net performance, then forecast runoff for existing commitments only.
- **Future Commitment Pacing Overlay.** Use user-provided planned commitments and strategy assumptions to estimate future calls, distributions, NAV, unfunded, exposure, and net cashflow. These values are planning assumptions, not Chronograph actuals.
- **Combined Liquidity Plan.** Add the existing portfolio forecast and pacing overlay to show total expected calls, distributions, net cashflow, NAV, unfunded, and exposure.

## Chronograph MCP Usage

The skill needs three categories of data from the Chronograph MCP server: entity identifiers (fund and group IDs), fund metadata (name, type, vintage, reporting currency, geographic focus, GP, final reporting date), and net LP-level cashflow values (NAV, 

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 3 — file references are metadata, not package contents.

View saved version →

---
name: chronograph-gp-meeting-prep
description: >
  Prepare an LP to meet with their fund manager (GP): review the fund's latest
  reporting, surface what changed since last period, and draft the questions
  worth raising. Use when someone is getting ready for a manager call, quarterly
  check-in, annual meeting, LPAC meeting, or a re-up decision — for example "I
  have my quarterly call with this manager next week, help me prep," "what
  changed in this fund this quarter," "what should I ask them about these marks,"
  or "what are the red flags in this reporting package." Draws on Chronograph
  fund and portfolio data when connected — captured reporting such as fund
  performance, schedules of investments, and portfolio company KPI profiles —
  and asks the LP to provide anything else it needs, such as a capital account
  statement, investor letter, or AGM or board deck.
---

# Chronograph GP Meeting Prep

**Requirements:** A connected Chronograph MCP server. These workflows are designed for permissioned Chronograph users to connect to their private investment data.

## Overview

Prepare an LP for a meeting with a GP by reviewing fund reporting, surfacing what changed, identifying diligence questions, and producing a concise meeting brief. The brief exists to sharpen a forward decision — a re-up, a secondary sale, or an LPAC vote — so prioritize what bears on those. Use Chronograph as the source of truth where connected and permissioned; otherwise work from user-provided reporting materials.

Default to a chat brief. Create a document, workbook, or deck only when the user asks for an artifact.

## Workflow

1. Clarify scope only when needed: GP, fund, reporting period, meeting type, as-of date, currency, units, output format, and whether the user wants a quick brief or a formal prep document.
2. Resolve the relevant fund, manager, commitment, reporting period, and LP perspective from Chronograph where available. If Chronograph is unavailable, ask for the GP reporting package or a pasted/uploaded summary.
3. Gather the fund baseline, anchoring first on the figures the GP reported this period and citing the source document where available: gross performance, cost, realized and unrealized value, current valuation, portfolio holdings, prior-period values, and fund attributes. Carry net performance — NAV, calls, distributions, TVPI/DPI/RVPI, and net IRR — as the LP's headline return, clearly labeled gross versus net and dated.
4. Review document-backed reporting content. When Chronograph document semantic search is available, use it as a primary step to find the story behind the numbers: what drove calls and distributions, how the GP describes its marks, and the commentary on portfolio company KPIs.
5. If a document the meeting needs isn't in the connected context, ask the LP to provide it. Do not infer narrative, management explanations, or meeting context from structured data alone.
6. Compare current period to prior period: performance movement, cashflow activity, NAV and unfunded movement, valuation changes, portfolio company adds/exits, concentration changes, and KPI movements.
7. Separate facts from interpretation. Use reporting data for facts, cite document support when available, and label inferred questions as analyst prompts rather than confirmed issues.
8. Produce a prioritized meeting brief with key changes, watch items, suggested GP questions, and follow-up requests.

## Chronograph Data Use

When the Chronograph MCP is available, use it to retrieve the relevant portfolio, fund, commitment, performance, exposure, and report-derived document context where permissioned. Keep public-facing explanations at the workflow level; do not expose internal schemas, private tool names, field mappings, or retrieval recipes.

Source discipline:

- Cite Chronograph, document, page, period, or user-provided source for quantitative claims whenever available.
- State currency, units, as-of date, and whether values are GP-reported, Ch

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 4 — file references are metadata, not package contents.

View saved version →

---
name: chronograph-portfolio-company-one-pager
description: >
  GP platform one-pager and investor report generator for private equity portfolio companies.
  Use this skill whenever a user asks to generate a company tearsheet, one-pager, investor
  report, portfolio overview, or company deep-dive — especially when they name a company or
  ask to "build a report", "create a one-pager", or "show me a tearsheet". Also trigger when
  the user asks to include commentary, quarterly updates, investment narratives, or any
  Investment Overview in the report output. This skill handles live data fetching via a
  connected MCP data source OR from an uploaded Excel model, metric formatting, AI-generated
  or model-sourced commentary, and rendering a fully styled HTML one-pager. Also trigger for
  LP quarterly updates, valuation summaries, and portco performance pages — any output that
  combines financials, valuation, and return data for a single portfolio company.
---

# GP Report Builder

Generates a fully styled, self-contained HTML investor report for a named portfolio company.
Supports two data source modes, two report types, and automatic brand detection from uploaded
templates. **Read this skill fully before writing a single line of HTML.**

**Requirements:** A connected Chronograph MCP server. These workflows are designed for permissioned Chronograph users to connect to their private investment data, or an uploaded Excel model for Model mode.

---

## Step 0 — Resolve Branding

Brand tokens are resolved automatically — the user never fills in a config file manually.
Execute the following decision tree before any other step.

### 0A — Check for an Uploaded Brand Template

Look for any of the following in the current conversation:

| File type | What to extract |
|---|---|
| **HTML / CSS file** | Parse all `color`, `background`, `font-family`, and `border` declarations. Identify the dominant background, primary accent, heading font, and body font. Extract any `--variable` tokens if a design system is present. |
| **PDF report or one-pager** | Visually analyse the document. Identify background color(s), the dominant accent/highlight color, heading and body typefaces, logo presence, and footer text. |
| **Image (PNG / JPG / SVG)** | Extract the 3–5 most visually prominent colors using the image content. Identify any visible text to infer font style (serif vs sans-serif, bold vs light). Note logo or wordmark if present. |
| **PowerPoint / PPTX** | Read slide backgrounds, title font, body font, accent colors from shapes and highlights. |
| **CSS / design token file** | Map token names to the brand token schema below directly. |

Once extracted, map findings to the Brand Token Schema in **0C**.

If the user has dropped **multiple** files, treat the most recent one as authoritative for
branding, unless the user specifies otherwise.

### 0B — No Template Provided → Apply Chronograph Defaults

If no brand template is present in the conversation, apply the following defaults silently —
do not ask the user to provide branding.

```
firm_name:             Chronograph
website:               www.chronograph.pe
confidentiality_label: CONFIDENTIAL

colors:
  bg_primary:       #101C1D   ← Near Black 1 (page background)
  bg_secondary:     #1A2627   ← Near Black 2 (card / panel background)
  bg_header:        #1B4147   ← Deep Teal (header strip)
  accent_primary:   #57E5EE   ← Bright Teal (headlines, KPI values, eyebrows)
  accent_secondary: #11A8B2   ← Regular Teal (bars, borders, left-rule accents)
  accent_negative:  #F95532   ← Accent Red (EBITDA bars, negative deltas)
  accent_positive:  #4ecb8a   ← Green (positive deltas)
  text_primary:     #FFFFFF   ← White (body text)
  text_muted:       #D9E8E8   ← Light teal tint (footnotes, commentary)
  table_header_bg:  #1B4147   ← Deep Teal (table header row)

fonts:
  heading:         Ubuntu
  heading_weight:  700
  body:            Open Sans
  body_weight:     300
  google_fonts_url: https://font

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 1 — file references are metadata, not package contents.