desk-setup
Install and verify the option desk command line tools so the option_* MCP tools and the optiondesk commands actually work. Use when a desk skill reports that no fresh figures can be produced, when optiondesk is not found, when the MCP tools are missing or failing, when the plugin was installed but nothing runs, or when the user asks how to set this up. Not for analysing options; it only gets the tools working.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- desk-setup
- description
- Install and verify the option desk command line tools so the option_* MCP tools and the optiondesk commands actually work. Use when a desk skill reports that no fresh figures can be produced, when optiondesk is not found, when the MCP tools are missing or failing, when the plugin was installed but nothing runs, or when the user asks how to set this up. Not for analysing options; it only gets the tools working.
03 SKILL.md 正文
Getting the desk working
The skills in this plugin describe a set of local tools. Installing the plugin gives you the instructions; it does not install the tools. This skill closes that gap.
First, find out what is actually missing
optiondesk doctor
If that runs, read its output: it reports whether the analytics engine is present, which providers can answer, which credentials are configured without printing them, and where artifacts are written. Fix what it names and stop here.
If the shell says command not found, continue.
Install
curl -fsSL https://raw.githubusercontent.com/Iman/agent-driven-options-desk-and-skills/main/install.sh | bash
That creates a virtualenv at ~/.optiondesk, installs the analytics engine
and the shell, links optiondesk and optiondesk-mcp into ~/.local/bin,
copies the skills into both ~/.claude/skills and ~/.agents/skills, and
registers the MCP server with every agent runtime it finds. Re-running is
safe, and ./install.sh --uninstall removes only what it created.
Python 3.11 or newer is required. Yahoo needs no API key, but it remains
disabled until the user accepts its local personal-use boundary. Use
--accept-yahoo-terms only after the user reads Yahoo's terms. The flag does
not permit hosting, business use, public display, or redistribution.
The acknowledgement is recorded as one line in ~/.optiondesk/config.env,
and that is where the provider gate reads it from. Nothing needs exporting
into a shell profile. Remove the line to withdraw it.
Run the demo from a checkout
This workflow installs the tools and runs analytics, including a paper position. Use it when the user requests the demo. For installation alone, use the installer described above.
If Yahoo access is already acknowledged, run:
./run.sh
For a first run, obtain the user's acknowledgement after they read Yahoo's terms.
Then run ./run.sh --accept-yahoo-terms.
Without that flag, the runner leaves acknowledgement to the installer's interactive prompt or the existing configuration.
Never add the flag merely because the user asked to install the tools.
From a checkout, this installs anything missing, repairs a virtualenv built
for the wrong CPU architecture, pulls live chains for SPY and QQQ, runs the
Greek ladder, dealer positioning, the structure comparison, a simulation and
a backtest per structure, opens a paper position, then serves the dashboard
and opens a browser. ./run.sh --help lists the options; --dry-run prints
every command without running one.
When the tools are installed but nothing imports
On Apple silicon, a script whose shebang is /usr/bin/env bash picks up a
Homebrew bash under /usr/local when one is installed, and that binary is
x86_64. Everything it installs is x86_64 too, so the virtualenv imports
cleanly from that shell and fails from every native one:
ImportError: ... incompatible architecture (have 'x86_64', need 'arm64')
The installer refuses to build under translation now and repairs wheels that
are already wrong. If a user is stuck on this before updating, re-running the
installer fixes it, and arch -arm64 ./install.sh forces the point.
The failure almost everyone hits
~/.local/bin is not on the PATH. The installer creates the two commands
there and warns rather than editing your shell profile, because editing
somebody's profile without asking is worse than a warning.
Check:
echo $PATH | tr ':' '\n' | grep -c "$HOME/.local/bin"
If that prints 0, add it to your shell profile:
export PATH="$HOME/.local/bin:$PATH"
Then open a new shell. This also matters to the MCP server, whose plugin
entry names the bare command optiondesk-mcp and resolves it through PATH.
Verify, in this order
optiondesk doctor engine, providers, credentials, paths
optiondesk expiries SPY reaches the network and lists expiries
optiondesk chain SPY writes the first artifact
optiondesk greeks --band 0.06 computes from it
If doctor works and expiries does not, the problem is network or
provider, not installation. If chain works and greeks reports the
engine is unavailable, the engine did not install; re-run the installer
without --no-engine.
What to tell the user when it still will not work
Say plainly that no fresh figures can be produced, and why. Do not substitute remembered or example numbers for a chain that was never pulled, and do not present a model value as a quote. A skill with no tools is a skill that explains the desk; it is not a desk.
What this cannot fix
In ChatGPT on the web or on a phone, there is no local process to install and nothing here will make local tools run. Do not ask the user to install a local process in that session. Use an attached remote MCP server when one is present. Without one, the skills can analyse user-provided data only.
Hosted deployments must set PUBLIC_DATA_MODE=demo or licensed. Demo mode
blocks all external providers. Licensed mode permits only a provider that has
an explicit approval for public display, derived outputs, storage, and MCP
delivery. Never change this mode to work around a provider refusal.
Reporting rules
They apply here too, briefly. Never present a remembered or example figure as a fresh one, and say plainly when the tools are absent rather than working around it. Never recommend a trade.
The full terms are in DISCLAIMER.md, which ships beside this skill when it is installed from a package and sits at the repository root otherwise.
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.