comps — Claude Code skill: comparables and installation
This Claude Code skill builds a trading comparables table from filings data. A request for company comps starts the workflow, which uses user-supplied share prices or market caps.
MIT ★ 86 updated 2026-11-16
/plugin marketplace add joaquimmj/deskstack
/plugin install deskstack@deskstackWhat comps does
The skill compares a target company with public peers and builds an EV/Revenue, EV/EBITDA, and P/E table. It gets revenue, EBITDA, net income, and net debt from deskstack filings data. EBITDA is calculated as operating income plus depreciation and amortization. If no peers are supplied, the skill proposes listed companies in the same industry and of similar size, then asks for confirmation.
Share prices with dates or market caps are needed for valuation; prices come from the user and are never estimated. If no price is supplied, financials stay in the table while EV and multiples appear as “n.a.”. The build uses the comps specification and build script from the analyst skill, followed by a self-check and offers to run /associate and /vp-review. EU issuers have annual figures only.
What it helps with
- compare a company with public peers using trading multiples
- gather peer financials for a selected reporting period
- calculate EV and multiples from user-supplied share prices
- check peer figures before handing off for review
Example prompts for the agent
- Build comps for ACME using MSFT and ORCL as peers.
- Find public peers for ACME and show me the proposed list to confirm.
- Create an EV/Revenue, EV/EBITDA, and P/E table for ACME using a share price of 120 USD dated 2025-03-01.
How to install comps
1Plugin marketplace
/plugin marketplace add joaquimmj/deskstack/plugin install deskstack@deskstackThis installs the whole deskstack plugin with all its skills
2Skills CLI (Vercel)
npx skills add joaquimmj/deskstack --skill comps -a claude-code3Manually: put the files in place
~/.claude/skills/comps/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/comps/
1Skills CLI (Vercel)
npx skills add joaquimmj/deskstack --skill comps -a codex2Manually: put the files in place
.agents/skills/comps/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add joaquimmj/deskstack --skill comps -a cursor2Manually: put the files in place
.cursor/skills/comps/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install joaquimmj/deskstack plugins/deskstack/skills/comps2Manually: put the files in place
.github/skills/comps/SKILL.md1Skills CLI (Vercel)
npx skills add joaquimmj/deskstack --skill comps -a gemini-cli2Platform CLI
gemini skills install https://github.com/joaquimmj/deskstack --path plugins/deskstack/skills/compsManually: put the files in place
.devin/skills/comps/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about comps
How do I install comps in Claude Code?
Run these commands in order: /plugin marketplace add joaquimmj/deskstack, then /plugin install deskstack@deskstack.
What does comps do?
It gathers financials for a target company and its peers, then builds an EV/Revenue, EV/EBITDA, and P/E table. Share prices must come from the user; the skill does not estimate them.
What does comps need to run?
It needs deskstack filings data, user-supplied share prices with dates or market caps, and the analyst skill with its comps specification and build script.
Source text
SKILL.md · MIT
---
name: comps
description: Builds a trading comparables table (EV/Revenue, EV/EBITDA, P/E) from deskstack filings data, with user-provided prices cited. Then suggests /associate.
argument-hint: "<company> [peer tickers]"
---
# Comps
1. **Target:** `company_lookup` for the target.
2. **Peers:**
- Use the user's peers if given.
- Otherwise propose 4–6 listed companies in the same industry and of similar size, and **ask the user to confirm
the list** (no peer tool yet).
3. **Financials:** for each company, `financial_statements` for the chosen period (LTM where available, else the
last fiscal year): revenue, EBITDA (computed: operating income + D&A), net income, net debt.
- EU issuers have annual figures only.
- Record each figure's own period in the spec (`"period": "FY2025"`), so the tie-out checks the right year.
4. **Prices:** ask the user for each share price and its date, or the market caps. Market cap = price × diluted
shares. Cite the price as "User-provided, <date>" and the shares to the filing. If the user gives no price, write
`"market_cap": {"value": null, "source": "Not given"}`: the row keeps its financials, and EV and the multiples show
"n.a.". Never estimate a price.
5. **Build:** follow the `analyst` skill's comps spec (`references/specs.md`) and run its build script. Keep each
company in its own reporting currency; multiples are currency-neutral.
6. **Hand off:** run the self-check, then offer `/associate` (ties out every peer's figures) and `/vp-review`.
End with "Draft for human review. Not investment advice."
© Joaquim · MIT · joaquimmj/deskstack