moss-trade-bot-factory-en-1.0.3

Name: moss-trade-bot-factory-en-1.0.3
Author: moss-site

380stars35forks

How to install moss-trade-bot-factory-en-1.0.3?

moss-site/moss-trade-bot-skills/moss-trade-bot-factory-en-1.0.3

$npx -y skills add moss-site/moss-trade-bot-skills --skill moss-trade-bot-factory-en-1.0.3

Installs into the current project.

Summary

Users describe a trading style in natural language, and the skill creates a crypto trading bot, runs local backtests, supports periodic reflection-driven evolution, and can optionally connect to an external platform for verification and simulated live trading.

Files · 1

View on GitHub

SKILL.md

1	# Moss Trade Bot Factory
2
3	A professional crypto quantitative trading bot factory plus strategy tuning specialist.
4
5	Knowledge base (read on demand, never all at once):
6	- Parameter reference + tuning cheatsheet -> `cat {baseDir}/knowledge/params_reference.md`
7	- Evolution logic + 7 reflection principles -> `cat {baseDir}/knowledge/evolution_guide.md`
8	- Verification upload + live trading operations -> `cat {baseDir}/knowledge/platform_ops.md`
9
10	## One-Time Local Setup
11
12	- This skill depends on the bundled `scripts/requirements.txt` for local Python packages such as `pandas`, `numpy`, `ccxt`, and `scipy`
13	- Install them only into the local `{baseDir}/.venv`; do not install extra packages and do not modify system-wide Python
14	- Before the first dependency install, ask one confirmation question such as `This skill needs its bundled local Python environment. Install it now into {baseDir}/.venv?`
15	- After the user approves, run:
16	```bash
17	cd {baseDir} && python3 scripts/setup_env.py
18	```
19	- After setup, prefer this interpreter for every bundled script:
20	```bash
21	PYTHON_BIN="{baseDir}/.venv/bin/python"
22	```
23	- If the current environment already imports all required packages successfully, you may continue without reinstalling them
24
25	## Safety And Transparency
26
27	- Local first: bot creation, backtesting, and evolution are local by default. If the user provides a CSV directly, the workflow can be fully offline
28	- Data boundary: backtest / evolution / verification upload always use Binance USDT-M or your provided Binance CSV. Coinbase is allowed only as a live signal input
29	- Platform features are optional: connect to the external platform only when the user explicitly asks to upload, bind, or trade live. The default platform URL comes from skill config `trade_api_url`, whose default is `https://ai.moss.site`
30	- Platform URL rule: `--platform-url` should be only the site origin, for example `https://ai.moss.site`; scripts will append the full API prefix and request `https://ai.moss.site/api/v1/moss/agent/agents/bind`
31	- Local credentials: platform credentials are stored at `~/.moss-trade-bot/agent_creds.json` by default; if skill config `agent_creds_path` is set, prefer that path. Credentials are sent only to the user-specified platform address
32	- No env vars: platform scripts depend only on explicit `--platform-url` and the local creds file. They do not read hidden environment variables and do not scan unrelated system credentials
33	- Explicit install mechanism: Python dependencies are installed only from the bundled `scripts/requirements.txt`, through `scripts/setup_env.py`, into the local `.venv`
34	- Progressive disclosure: multiple local markdown files are read only when needed; `/tmp/*.json` files are only local intermediates for params, fingerprints, and backtest outputs
35	- Confirmation boundary: stop at these checkpoints: whether to enable weekly evolution, one confirmation before the first dependency install, one execution confirmation before the first local backtest run, the A/B/C choice after backtest results, the first live data-source switch, and each order in manual mode. All other local steps should continue directly
36
37	Follow the steps below strictly. Do not skip steps. Stop only at the confirmation checkpoints explicitly called out below; continue directly for everything else.
38
39	---
40
41	## Step 1: Understand Intent And Confirm Evolution
42
43	After receiving the strategy description, use professional judgment to auto-fill the configuration. Ask only one question: whether to enable evolution.
44
45	Automatic inference rules (do not ask item by item):
46	- Direction: trend-following -> bi-directional (`0.5`), bearish / contrarian -> short-biased (`0.1~0.3`), conservative / DCA -> long-biased (`0.6~0.8`)
47	- Leverage: conservative -> `3~5x`, neutral -> `8~12x`, aggressive -> `15~25x`, all-in -> `50~100x`
48	- Defaults: `BTC/USDT`, `15m`, `148 days`, `$10,000`
49
50	You must ask the user:
51	```text
52	Do you want to enable weekly evolution?
53	On: every week, tactical parameters are fine-tuned based on trading results while the core personality stays the same. Best for trend / momentum strategies
54	Off: parameters stay fully fixed. Best for highly disciplined strategies or when you already trust the setup
55	Default recommendation: On
56	```
57
58	Backtest data prerequisite: you must have an OHLCV CSV before running a backtest.
59
60	- Local experimentation: any Binance UM time range is acceptable
61	- Verification upload: you must use the 2025-10-06 ~ 2026-03-03 range. `fetch_data.py` defaults to this

People who liked this also viewed

6 picks

Type

moss-trade-bot-factory-en-1.0.3

380stars35forks

View on GitHub

How to install moss-trade-bot-factory-en-1.0.3?

moss-site/moss-trade-bot-skills/moss-trade-bot-factory-en-1.0.3

$npx -y skills add moss-site/moss-trade-bot-skills --skill moss-trade-bot-factory-en-1.0.3

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Moss Trade Bot Factory
2
3	A professional crypto quantitative trading bot factory plus strategy tuning specialist.
4
5	Knowledge base (read on demand, never all at once):
6	- Parameter reference + tuning cheatsheet -> `cat {baseDir}/knowledge/params_reference.md`
7	- Evolution logic + 7 reflection principles -> `cat {baseDir}/knowledge/evolution_guide.md`
8	- Verification upload + live trading operations -> `cat {baseDir}/knowledge/platform_ops.md`
9
10	## One-Time Local Setup
11
12	- This skill depends on the bundled `scripts/requirements.txt` for local Python packages such as `pandas`, `numpy`, `ccxt`, and `scipy`
13	- Install them only into the local `{baseDir}/.venv`; do not install extra packages and do not modify system-wide Python
14	- Before the first dependency install, ask one confirmation question such as `This skill needs its bundled local Python environment. Install it now into {baseDir}/.venv?`
15	- After the user approves, run:
16	```bash
17	cd {baseDir} && python3 scripts/setup_env.py
18	```
19	- After setup, prefer this interpreter for every bundled script:
20	```bash
21	PYTHON_BIN="{baseDir}/.venv/bin/python"
22	```
23	- If the current environment already imports all required packages successfully, you may continue without reinstalling them
24
25	## Safety And Transparency
26
27	- Local first: bot creation, backtesting, and evolution are local by default. If the user provides a CSV directly, the workflow can be fully offline
28	- Data boundary: backtest / evolution / verification upload always use Binance USDT-M or your provided Binance CSV. Coinbase is allowed only as a live signal input
29	- Platform features are optional: connect to the external platform only when the user explicitly asks to upload, bind, or trade live. The default platform URL comes from skill config `trade_api_url`, whose default is `https://ai.moss.site`
30	- Platform URL rule: `--platform-url` should be only the site origin, for example `https://ai.moss.site`; scripts will append the full API prefix and request `https://ai.moss.site/api/v1/moss/agent/agents/bind`
31	- Local credentials: platform credentials are stored at `~/.moss-trade-bot/agent_creds.json` by default; if skill config `agent_creds_path` is set, prefer that path. Credentials are sent only to the user-specified platform address
32	- No env vars: platform scripts depend only on explicit `--platform-url` and the local creds file. They do not read hidden environment variables and do not scan unrelated system credentials
33	- Explicit install mechanism: Python dependencies are installed only from the bundled `scripts/requirements.txt`, through `scripts/setup_env.py`, into the local `.venv`
34	- Progressive disclosure: multiple local markdown files are read only when needed; `/tmp/*.json` files are only local intermediates for params, fingerprints, and backtest outputs
35	- Confirmation boundary: stop at these checkpoints: whether to enable weekly evolution, one confirmation before the first dependency install, one execution confirmation before the first local backtest run, the A/B/C choice after backtest results, the first live data-source switch, and each order in manual mode. All other local steps should continue directly
36
37	Follow the steps below strictly. Do not skip steps. Stop only at the confirmation checkpoints explicitly called out below; continue directly for everything else.
38
39	---
40
41	## Step 1: Understand Intent And Confirm Evolution
42
43	After receiving the strategy description, use professional judgment to auto-fill the configuration. Ask only one question: whether to enable evolution.
44
45	Automatic inference rules (do not ask item by item):
46	- Direction: trend-following -> bi-directional (`0.5`), bearish / contrarian -> short-biased (`0.1~0.3`), conservative / DCA -> long-biased (`0.6~0.8`)
47	- Leverage: conservative -> `3~5x`, neutral -> `8~12x`, aggressive -> `15~25x`, all-in -> `50~100x`
48	- Defaults: `BTC/USDT`, `15m`, `148 days`, `$10,000`
49
50	You must ask the user:
51	```text
52	Do you want to enable weekly evolution?
53	On: every week, tactical parameters are fine-tuned based on trading results while the core personality stays the same. Best for trend / momentum strategies
54	Off: parameters stay fully fixed. Best for highly disciplined strategies or when you already trust the setup
55	Default recommendation: On
56	```
57
58	Backtest data prerequisite: you must have an OHLCV CSV before running a backtest.
59
60	- Local experimentation: any Binance UM time range is acceptable
61	- Verification upload: you must use the 2025-10-06 ~ 2026-03-03 range. `fetch_data.py` defaults to this

People who liked this also viewed

6 picks

Type