<!-- carta:instrumentation-fallback --> <IMPORTANT> _instrumentation_v2 is anonymous Carta skill-usage telemetry. ALWAYS include it in your Carta MCP tool-call arguments — the plugins' hooks overwrite it when they run, so always sending it is safe. Shape: _instrumentation_v2: {"skills": ["<plugin>:<skill>", ...], "model": "<model-id>", "surface": "<value>"} skills is cumulative for the session: list every Carta skill loaded so far, not only the one driving this call. Namespace each entry "plugin:skill" and keep them oldest-first, with the most recently used skill last. List only Carta skills (carta-cap-table:*, carta-crm:*, carta-investors:*) — never skills from non-Carta plugins. surface is the Claude surface you are running in: "chat" (claude.ai or the Claude app, i.e. regular chat, not Cowork), "cowork" (Cowork mode), "code-terminal", "code-desktop", or "excel". Omit it entirely if none of those describe your surface or you cannot tell — do not guess and do not invent another value. </IMPORTANT>

<!-- Part of the official Carta AI Agent Plugin -->

Market Benchmarks

Compute portfolio-wide benchmarks from your own Carta data: option pool sizes, SAFE valuation caps, and round sizes. Useful for sanity-checking a new deal's terms against your existing portfolio.

Note: This reflects your firm's portfolio, not Carta-wide market data. Present results as "portfolio benchmarks" not "market data."

Prerequisites

No inputs required — this skill loops the full portfolio automatically.

Data Retrieval

Portfolio Enumeration

Call list_accounts. Filter to corporation_pk: accounts. Extract up to 20 numeric corporation IDs. If more than 20 companies exist, ask the user to narrow scope.

Per-Company Commands

For each company, the relevant commands are:

• call_tool({"name": "cap_table__get__cap_table_by_share_class", "arguments": {"corporation_id": corporation_id}}) -- option pool data
• call_tool({"name": "cap_table__get__convertible_notes", "arguments": {"corporation_id": corporation_id}}) -- SAFE/note terms (summary includes median/min/max price_cap, avg_discount, by_type)
• call_tool({"name": "cap_table__get__financing_history", "arguments": {"corporation_id": corporation_id}}) -- round sizes (summary includes per-round cash_raised and latest_date)

The gateway defaults to detail=summary for all three commands. The enriched summaries include all fields needed for portfolio benchmarks — no individual records required.

Parallel execution: The fetch tool has readOnlyHint=true, so Claude Code executes parallel fetch calls concurrently. Issue ALL fetch calls for ALL companies in a single response — do NOT loop company-by-company. See Workflow Step 2.

Key Fields

From cap table (option pool):

• option_plans[].authorized_shares: shares authorized per plan
• totals.total_fully_diluted: total fully diluted share count

From convertible notes (summary):

• median_price_cap, min_price_cap, max_price_cap: valuation cap statistics
• avg_discount: average discount rate
• by_type: count of SAFEs vs Convertible Notes
• total_dollar_amount: total invested across all instruments

From financing history (summary):

• by_round: per-round {count, cash_raised, latest_date}
• total_cash_raised: aggregate across all rounds

Workflow

Step 1 — Get Portfolio

Call list_accounts. Filter to corporation_pk: accounts. Extract up to 20 numeric corporation IDs.

Step 2 — Collect Data for All Companies (parallel)

Issue ALL fetch calls for ALL companies in a single response — do NOT loop company-by-company. Each fetch call is independent and will execute concurrently.

For example, with 5 companies and all 3 data types, issue all 15 fetch calls at once:

call_tool({"name": "cap_table__get__cap_table_by_share_class", "arguments": {"corporation_id": 1}})
call_tool({"name": "cap_table__get__convertible_notes", "arguments": {"corporation_id": 1}})
call_tool({"name": "cap_table__get__financing_history", "arguments": {"corporation_id": 1}})
call_tool({"name": "cap_table__get__cap_table_by_share_class", "arguments": {"corporation_id": 2}})
call_tool({"name": "cap_table__get__convertible_notes", "arguments": {"corporation_id": 2}})
call_tool({"name": "cap_table__get__financing_history", "arguments": {"corporation_id": 2}})
... (all companies)

Then from the results:

Cap table by share class (for option pool %):

• From option_plans[]: sum authorized_shares across all plans
• From totals.total_fully_diluted: compute option pool % = option_pool_authorized / total_fully_diluted

SAFE / convertible note terms (summary):

• Use median_price_cap, min_price_cap, max_price_cap directly for SAFE cap benchmarks
• Use avg_discount for discount benchmarks
• Use by_type to count SAFEs vs notes per company

Financing history (summary):

• Use by_round to identify rounds and their cash_raised
• Use total_cash_raised for aggregate amounts
• Most recent round = round with latest latest_date

Step 3 — Compute Summary Statistics

For each metric, compute across companies that have data:

• Median, min, max
• Skip companies with no data for a given metric (don't count as zero)

Metrics:

• Option pool % (fully diluted)
• SAFE valuation cap
• Last priced round size

Step 4 — Present Results

See Presentation section.

If the user asks about a specific company ("how does Acme's option pool compare?"), show that company's value alongside the portfolio median.

Gates

Required inputs: None — portfolio enumeration is automatic.

AI computation: Yes — portfolio benchmark statistics (median, min, max for option pool sizes, SAFE caps, round sizes) are AI-derived from aggregated cap table data. Trigger the AI computation gate (see carta-interaction-reference §6.2) before outputting any benchmark statistics or portfolio comparisons.

Subagent prohibition: Not applicable.

Presentation

Format: Benchmark tables grouped by metric

BLUF lead: Lead with the number of companies analyzed and the most notable finding (e.g., "median option pool is 12.5% across 14 companies").

Sort order: By metric name (Option Pool, SAFE Caps, Round Sizes).

Portfolio Benchmarks (N companies)

Option Pool Size (% Fully Diluted)

SAFE Valuation Caps

Last Priced Round Size

Caveats

• Portfolio data reflects point-in-time API calls, not a single atomic snapshot
• Companies with restricted permissions may have incomplete data
• Rate limit: maximum 20 companies per invocation
• This reflects your firm's portfolio, not Carta-wide market data — present results as "portfolio benchmarks" not "market data"