/docs/ai-assistants
Use with your AI assistant
Six prompts to copy into your AI coding assistant.
Copy a prompt into your AI coding assistant (Claude, Codex, Cursor, Copilot, Windsurf, Goose, Antigravity and others). Each one points the assistant at the right page and at the same ground rules.
Ground rules in every prompt
- Read /llms.txt and the relevant docs page first.
- Key from ERS_API_KEY, never in the code.
- Check only before opening a position, never before closing one; the strategy stays as it is.
- You decide the UNKNOWN policy (allow or block); the assistant asks you.
- At most one request per 5 minutes; Retry-After respected; a short test included.
- A code comment that this is a risk check, not a trade instruction.
Python bot with ERS Guard
Add an entry-risk check to my Python trading bot with ERS Guard (pip install git+https://github.com/entryriskscore/ers-guard). Before each new entry, use guard.allow_entry(symbol, side, hold="60m", block={"HIGH"}) and skip the entry when it returns False. First read https://entryriskscore.com/llms.txt and https://entryriskscore.com/docs/guard. Read the API key from the environment variable ERS_API_KEY; never hard-code it. Check only before opening a position, never before closing one. Keep my strategy logic unchanged. Ask me what the bot should do when the answer is UNKNOWN (allow or block). Ask the API at most once per 5 minutes and respect Retry-After. Add a short test. Add a code comment saying this is a risk check, not a trade instruction.Freqtrade strategy
Add ERS Guard to my Freqtrade strategy: copy integrations/freqtrade/ers_guard_mixin.py from the ERS Guard repository and make my strategy class inherit ErsGuardMixin before IStrategy, with ers_block = {"HIGH"} and ers_hold = "60m". First read https://entryriskscore.com/llms.txt and https://entryriskscore.com/docs/guard#freqtrade. Read the API key from the environment variable ERS_API_KEY; never hard-code it. Check only before opening a position, never before closing one. Keep my strategy logic unchanged. Ask me what the bot should do when the answer is UNKNOWN (allow or block). Ask the API at most once per 5 minutes and respect Retry-After. Add a short test. Add a code comment saying this is a risk check, not a trade instruction.Node.js bot
Add an entry-risk check to my Node.js bot with js/ers-guard.js from the ERS Guard repository (Node 18+, no dependencies). Before each new entry, await guard.allowEntry(symbol, side, { block: ["HIGH"] }). First read https://entryriskscore.com/llms.txt and https://entryriskscore.com/docs/guard#javascript. Read the API key from the environment variable ERS_API_KEY; never hard-code it. Check only before opening a position, never before closing one. Keep my strategy logic unchanged. Ask me what the bot should do when the answer is UNKNOWN (allow or block). Ask the API at most once per 5 minutes and respect Retry-After. Add a short test. Add a code comment saying this is a risk check, not a trade instruction.Any language via the sidecar
My bot is written in another language. Run "ers-guard serve --port 8787" as a sidecar and, before each new entry, read http://127.0.0.1:8787/check?symbol=SYMBOL&side=SIDE&hold=60m; skip the entry when "level" is "HIGH". First read https://entryriskscore.com/llms.txt and https://entryriskscore.com/docs/guard#sidecar. Read the API key from the environment variable ERS_API_KEY; never hard-code it. Check only before opening a position, never before closing one. Keep my strategy logic unchanged. Ask me what the bot should do when the answer is UNKNOWN (allow or block). Ask the API at most once per 5 minutes and respect Retry-After. Add a short test. Add a code comment saying this is a risk check, not a trade instruction.
MCP set-up in my IDE
Connect my IDE's AI assistant to the Entry Risk Score MCP server (https://entryriskscore.com/api/v1/mcp, header Authorization: Bearer with the key from ERS_API_KEY) using the exact configuration for my client from the docs, then list the four tools and show me which coin/sides are HIGH right now. First read https://entryriskscore.com/llms.txt and https://entryriskscore.com/docs/mcp. Read the API key from the environment variable ERS_API_KEY; never hard-code it. Check only before opening a position, never before closing one. Keep my strategy logic unchanged. Ask me what the bot should do when the answer is UNKNOWN (allow or block). Ask the API at most once per 5 minutes and respect Retry-After. Add a short test. Add a code comment saying this is a risk check, not a trade instruction.
Plain REST from the OpenAPI document
Write a small client for the Entry Risk Score REST API from https://entryriskscore.com/api/v1/openapi.json. Use GET /api/v1/state with X-API-Key, cache the answer until meta.next_update_at, and expose a function entry_risk(symbol, side, profile) that returns status, score and as_of for my bot to use before new entries. First read https://entryriskscore.com/llms.txt and https://entryriskscore.com/docs/api. Read the API key from the environment variable ERS_API_KEY; never hard-code it. Check only before opening a position, never before closing one. Keep my strategy logic unchanged. Ask me what the bot should do when the answer is UNKNOWN (allow or block). Ask the API at most once per 5 minutes and respect Retry-After. Add a short test. Add a code comment saying this is a risk check, not a trade instruction.