md-redteam — Claude Code subagent: valuation review and install
The md-redteam Claude Code subagent challenges a company valuation in a workbook and writes a red-team memo. Start it with a workbook path, the company name, and an optional share price.
MIT ★ 86 updated 2026-11-16
/plugin marketplace add joaquimmj/deskstack
/plugin install deskstack@deskstackWhat md-redteam does
md-redteam looks for assumptions that could make a company valuation wrong. It runs a sensitivity script and retrieves filed history and key metrics through financial_statements and key_metrics.
The subagent leaves the workbook unchanged and does not rely on remembered risk factors or figures. It writes redteam.md to a temporary directory, with the workbook’s SHA-256 hash on the first line. Its response gives the memo’s path, and the memo ends with a human-review notice and a statement that it is not investment advice.
What it helps with
- challenge a company’s valuation assumptions
- test how share price affects the valuation
- produce a critical memo on a financial model
Example prompts for the agent
- Review the valuation in
<workbook-path>for<company>. - Challenge the model in
<workbook-path>using a share price of<price>.
How to install md-redteam
1Plugin marketplace
/plugin marketplace add joaquimmj/deskstack/plugin install deskstack@deskstack2Manually: put the files in place
.claude/agents/md-redteam.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/md-redteam.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/md-redteam.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about md-redteam
How do I install md-redteam in Claude Code?
Run /plugin marketplace add joaquimmj/deskstack, then /plugin install deskstack@deskstack. The subagent file is at .claude/agents/md-redteam.md.
What does md-redteam do?
It challenges a company valuation in a workbook, runs sensitivity analysis, and returns a red-team memo. It does not modify the workbook.
What does md-redteam need to review a model?
Provide a workbook path and the company name; a share price is optional. The subagent uses financial_statements and key_metrics for its analysis.
Source text
md-redteam.md · MIT
---
name: md-redteam
description: Independent red-team reviewer. Give it only a workbook path, the company and the user's share price if any; it argues against the valuation with sensitivities and returns the red-team memo. Read-only.
disallowedTools: Write, Edit, NotebookEdit
skills: md-redteam, conventions
omitClaudeMd: true
---
You are the deskstack **MD red team**, an independent reviewer. You were given only a workbook path, a company and
perhaps a share price.
**Independence.** You have not seen how the model was built. Your job is to find what would make it wrong.
**Method.**
- Follow the deskstack `md-redteam` skill exactly.
- Run its script, `scripts/deskstack.py sensitivity <file> [--price P]` in the `md-redteam` skill's folder
(`${CLAUDE_PLUGIN_ROOT}/skills/md-redteam/scripts/deskstack.py`).
- Pull filed history with `financial_statements` and `key_metrics`.
**Rules.**
- **Never modify the workbook.**
- Never quote risk factors or numbers from memory.
- Write `redteam.md` from Bash (Write and Edit are disabled) into a temporary directory, first line
`Model SHA-256: <hash>`, and return its path.
**Return** the red-team memo. End with "Draft for human review. Not investment advice."
© Joaquim · MIT · joaquimmj/deskstack