Alpaca Paper Live Key Separation logo

Alpaca Paper Live Key Separation

Community
HimanshuJ16
alpaca-paper-live-key-separation

Use when a bot connects to the Alpaca Trading API and must never cross paper keys with the live endpoint; pins the base URL per environment, checks the key prefix, probes /v2/account and requires an explicit ALLOW_LIVE_TRADING flag.

Overview

PublisherHimanshuJ16
RepositoryAlgo-Trading-Skills
Skill namealpaca-paper-live-key-separation
Stars
50
Forks
7
Bundled files
5
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 5 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by HimanshuJ16 on GitHub. Read the source before you install it.

Installation

Install the Alpaca Paper Live Key Separation AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/HimanshuJ16/Algo-Trading-Skills/tree/main/skills/alpaca-paper-live-key-separation
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/HimanshuJ16/Algo-Trading-Skills.git /tmp/Algo-Trading-Skills
mkdir -p .claude/skills
cp -r /tmp/Algo-Trading-Skills/skills/alpaca-paper-live-key-separation .claude/skills/alpaca-paper-live-key-separation
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Alpaca Paper Live Key Separation in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Alpaca Paper Live Key Separation on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Alpaca Paper Live Key Separation is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

When to Use

Invoke this whenever a bot connects to the Alpaca Trading API (or any broker with distinct paper/live environments). Connecting a paper strategy to Alpaca's live endpoint (https://api.alpaca.markets) using live API keys — or passing live order signals into a paper endpoint — is a catastrophic operational error. Pinning the base URL per environment, inspecting the key prefix, probing /v2/account for tradability, and requiring an explicit ALLOW_LIVE_TRADING environment variable are all mandatory before any order is submitted.

The base URL is the control that actually separates the environments. Alpaca serves paper accounts from https://paper-api.alpaca.markets and live accounts from https://api.alpaca.markets; a live account is not reachable through the paper host. Every other check in this skill is defence-in-depth layered on that pin — treat them as corroboration, never as a substitute.

When NOT to Use

  • Brokers without separate paper/live endpoints: If a single API base URL serves both environments and differentiation is done purely via credentials, this skill's URL-matching logic does not apply. Use broker-specific auth patterns instead (see headless-broker-auth-patterns).
  • Backtesting or simulation engines: When running historical replays or simulated fills with no network calls to a real broker API, environment segregation is irrelevant.
  • Non-Alpaca brokers with different credential schemes: The PK.../AK... prefix convention is Alpaca-specific. Brokers using different key formats (e.g., IBKR account IDs DU.../U...) require their own validation logic. For a broker-agnostic host allow-list, see sandbox-credential-leakage-prevention.
  • Read-only market data access: If the integration only consumes market data endpoints (not order routing), live capital loss is not a risk and the order guard is unnecessary.

Prerequisites

  • Distinct environment variable names for paper vs live credentials (e.g. ALPACA_PAPER_KEY_ID vs ALPACA_LIVE_KEY_ID).
  • Base URL configuration (https://paper-api.alpaca.markets for paper vs https://api.alpaca.markets for live).
  • Explicit ALLOW_LIVE_TRADING=true environment flag for live execution mode.

Workflow

  1. Load & Normalise the Environment Mode:

    • Coerce the configured environment into a known PAPER/LIVE value before any comparison. A value matching neither must raise, never fall through — an unrecognised mode that skips both branches is an authorisation, not a no-op.
  2. Inspect Credential Prefixes:

    • PK... indicates Alpaca paper credentials; AK... indicates live credentials.
    • Use this to reject a credential carrying the opposite environment's prefix. Do not require a positive prefix match: the convention is widely observed but is not documented by Alpaca, so an unrecognised key format must not be treated as proof of anything (see references/standards.md).
  3. Base URL & Mode Verification:

    • Match configuration mode against endpoint URLs:
      • PAPER mode → https://paper-api.alpaca.markets
      • LIVE mode → https://api.alpaca.markets
    • Match the exact URL against an allow-list, case-normalised. Anything else — including a look-alike host such as https://api.alpaca.markets.attacker.example — is rejected. Never use a substring or startswith test here.
  4. Live Execution Safety Gate (ALLOW_LIVE_TRADING):

    • Block initialization in LIVE mode unless ALLOW_LIVE_TRADING=true is explicitly set. Strip the value before comparing — a trailing newline from a .env loader should not silently block a legitimate live deployment.
    • Emit a WARNING when live trading is authorised, so the transition is visible in the log record.
  5. Account API Probe:

    • Issue GET /v2/account on startup and reject a response that is not a mapping.
    • Verify the account is tradable: status must be ACTIVE (or PAPER_ONLY, which is valid only in paper mode). A missing status is a veto, not an assumed ACTIVE.
    • Veto if any order-blocking flag is set: trading_blocked, account_blocked, or trade_suspended_by_user. Alpaca documents the first and last as "the account is not allowed to place orders" — a guard that ignores them authorises orders the broker will reject.
    • Resolve the environment only from signals that actually exist. GET /v2/account does not return an is_paper field (verified against Alpaca's account schema and the official alpaca-py TradeAccount model). Use, in order: an is_paper bool if your SDK wrapper injects one; status == "PAPER_ONLY"; an account_number beginning PA (observed, unofficial — treat as a positive paper signal only, never as proof an account is live).
    • If none of those resolve, the environment is undeterminable — log it and fall back on the already-verified base URL. Do not treat "undeterminable" as live: that vetoes every legitimate paper deployment, because the real API never supplies the field.
  6. Order Submission Veto Guard:

    • Wrap order routing calls in AlpacaEnvironmentManager.guard_order(), vetoing any outbound order if environment checks fail or key/URL mismatches occur.
    • Validate the order itself at the gate — reject a non-positive, non-finite, or non-numeric qty, an empty symbol, and any side outside buy/sell.

Full step-by-step procedure with broker-specific detail: see references/workflows.md. Broker/framework coverage table for this skill: see references/standards.md. Printable pre-flight checklist: see assets/checklist.md.

Common Pitfalls

  • Hardcoding Base URLs: Hardcoding https://api.alpaca.markets in code and relying only on switching API keys in .env.
  • Shared Credential Variable Names: Using generic ALPACA_KEY_ID for both paper and live testing, leading to accidental live deployment.
  • Trusting a nonexistent is_paper field: Alpaca's /v2/account response has no is_paper key. Code written against it reads None on every call — so a guard that treats a missing value as live bricks paper trading, and one that treats it as paper waves live accounts through. Resolve the environment from fields the API really returns, and treat "unknown" as unknown.
  • Letting an unrecognised environment fall through: if mode == PAPER … elif mode == LIVE … with no else returns success for any third value, so a typo'd or foreign enum member authorises a live order with no ALLOW_LIVE_TRADING check. Normalise the mode up front and raise on anything unknown.
  • Defaulting a missing status to ACTIVE: A truncated or error-shaped account payload then reads as a healthy account. An unreadable account is a veto.
  • Ignoring the blocked flags: trading_blocked / account_blocked / trade_suspended_by_user mean the broker will refuse the order. Catching that locally turns a confusing broker rejection into a clear pre-trade veto.
  • Strict is True on a loosely-typed payload: An adapter that hands back "true"/"false" as strings defeats if account_data.get("trading_blocked") is True, so an explicitly blocked account reads as unblocked. Accept the string forms and fail closed on a value you cannot interpret.
  • Substring URL matching: base_url.startswith("https://api.alpaca.markets") also accepts https://api.alpaca.markets.attacker.example. Compare against an exact allow-list.
  • Missing Live Confirmation Flag: Allowing live trading without an explicit boolean environment variable guard (ALLOW_LIVE_TRADING=true).

Verification

  • Configure paper keys with the live URL and confirm AlpacaEnvironmentManager raises EnvironmentMismatchError.
  • Attempt live mode initialization without ALLOW_LIVE_TRADING=true and confirm execution is blocked.
  • Construct a config with an unrecognised environment value and confirm it raises rather than validating.
  • Simulate an account probe returning is_paper=False when configured in paper mode and confirm the startup veto.
  • Simulate a realistic paper payload with no is_paper field (status=ACTIVE, account_number=PA…) and confirm it is accepted, not vetoed.
  • Simulate trading_blocked=true and confirm the order guard vetoes.
  • Simulate a probe response with no status field and confirm the veto.
  • Submit qty=0, qty=NaN, and side="long" through guard_order() and confirm each is rejected.
  • Run the unit test suite python -m unittest discover -s skills/alpaca-paper-live-key-separation/scripts and confirm a 100% pass rate.

Related Skills

  • paper-to-live-promotion-checklist
  • headless-broker-auth-patterns
  • kill-switch-and-drawdown-circuit-breakers
  • sandbox-credential-leakage-prevention
  • sandbox-vs-production-endpoint-drift

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Alpaca Paper Live Key Separation AI skill do?

Use when a bot connects to the Alpaca Trading API and must never cross paper keys with the live endpoint; pins the base URL per environment, checks the key prefix, probes /v2/account and requires an explicit ALLOW_LIVE_TRADING flag.

Why use Alpaca Paper Live Key Separation on TypingMind?

Because you install it once and use it with any model. Alpaca Paper Live Key Separation is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Alpaca Paper Live Key Separation in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/HimanshuJ16/Algo-Trading-Skills/tree/main/skills/alpaca-paper-live-key-separation. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Alpaca Paper Live Key Separation?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Alpaca Paper Live Key Separation?

As many as you like. As long as a model supports skills, you can use Alpaca Paper Live Key Separation with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Alpaca Paper Live Key Separation AI skill free?

Yes. It is published on GitHub by HimanshuJ16 under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇