← Back to Skills Marketplace
mibayy

Polymarket Clob Microstructure

by Mibayy · GitHub ↗ · v2.0.6 · MIT-0
cross-platform ✓ Security Clean
141
Downloads
0
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install clob-microstructure
Description
Analyzes Polymarket CLOB order books to score liquidity gaps, order imbalance, whale activity, and fake breakouts, generating trading signals and fading fake...
README (SKILL.md)

CLOB Microstructure Scanner

Structural alpha from Polymarket order book analysis. Nobody else on Simmer does this.

What It Does

Pulls raw order book data from Polymarket's CLOB API and scores each market on four microstructure dimensions:

Dimension What It Measures
Liquidity Gaps Thin zones in the book where price can jump
Order Book Imbalance Bid vs ask weight asymmetry
Whale Activity Large single orders relative to book depth
Fake Breakout Detection Sudden moves on thin volume that revert

Each dimension scores 0-25, combined into a 0-100 inefficiency score.

Signals

  • ENTRY_OK — Normal conditions, standard sizing
  • REDUCE_SIZE — Elevated microstructure risk, cut position size
  • SKIP — Too dangerous, stay out
  • MEAN_REVERT — Fake breakout detected, fade the move

The bot only trades on MEAN_REVERT signals, fading fake breakouts via SimmerClient.

Configuration

  • Runs every 5 minutes via cron
  • Defaults to dry-run; pass --live for real execution
  • Requires SIMMER_API_KEY environment variable

Usage

# Dry run (default)
python clob_microstructure.py

# Live trading
python clob_microstructure.py --live

# Quiet mode for cron
python clob_microstructure.py --live --quiet

🧪 Remixable Template — Fork this skill and tweak the scoring weights, signal thresholds, or add your own microstructure indicators. The modular scoring system makes it easy to extend.

Usage Guidance
This skill appears coherent with its described purpose. Before enabling live trading: 1) Keep it in dry-run to observe signals and log output; 2) Confirm the permissions and behaviour of the Simmer account tied to SIMMER_API_KEY (what trade permissions and withdrawal/transfer rights it has); 3) Ensure the default TRADING_VENUE='sim' or avoid --live until you trust results; 4) Review and tune MAX_TRADES_PER_RUN / FADE_SIZE_USD to limit financial exposure; 5) Be aware pip dependencies will be installed from PyPI (simmer-sdk, requests). If you need higher assurance, audit the full clob_microstructure.py file (it’s included) and verify SimmerClient’s implementation separately.
Capability Analysis
Type: OpenClaw Skill Name: clob-microstructure Version: 2.0.6 The skill is a legitimate trading bot designed to analyze Polymarket order book microstructure and execute 'mean reversion' trades via the Simmer SDK. The Python script (clob_microstructure.py) implements a multi-dimensional scoring system (liquidity gaps, imbalance, whale activity, and fake breakouts) and interacts with official Polymarket API endpoints. No evidence of data exfiltration, malicious execution, or prompt injection was found.
Capability Assessment
Purpose & Capability
Name/description match the code: it fetches order books and trades MEAN_REVERT signals via SimmerClient. Declared dependency on simmer-sdk and requests is consistent with this functionality; required env var SIMMER_API_KEY is exactly what a trading client needs.
Instruction Scope
SKILL.md and the Python entrypoint focus on: fetching CLOB and trade data from polymarket endpoints, scoring markets, safety checks, and sending trades through SimmerClient. The instructions do not ask the agent to read unrelated files, secrets, or system paths, nor to post data to unexpected external endpoints.
Install Mechanism
clawhub.json lists pip packages (simmer-sdk, requests) so installation will pull from PyPI — this is expected and moderate risk. There is no download-from-arbitrary-URL or archive extraction. (The top-level metadata said 'no install spec' but clawhub.json contains pip requirements — the presence of pip deps is normal.)
Credentials
Only SIMMER_API_KEY (and optional TRADING_VENUE and a few CLOB_* tuning envs) are referenced. These are proportional to a trading bot that uses SimmerClient; the skill does not request unrelated secrets or multiple external credentials.
Persistence & Privilege
always is false and the skill is user-invocable. It does not request global or elevated privileges, nor does it modify other skills' configuration. Autonomous invocation is allowed (default) but not combined with other red flags.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install clob-microstructure
  3. After installation, invoke the skill by name or use /clob-microstructure
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.6
Gradient sizing: 75-79=, 80-89=0, 90-94=5, 95+=0 — proportional to conviction score
v2.0.5
Expanded SKIP_PATTERNS: weather, political, social/influencer, long-horizon markets — no microstructure edge on these
v2.0.1
Fix Market dataclass access, find_markets API, venue default to sim
Metadata
Slug clob-microstructure
Version 2.0.6
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 3
Frequently Asked Questions

What is Polymarket Clob Microstructure?

Analyzes Polymarket CLOB order books to score liquidity gaps, order imbalance, whale activity, and fake breakouts, generating trading signals and fading fake... It is an AI Agent Skill for Claude Code / OpenClaw, with 141 downloads so far.

How do I install Polymarket Clob Microstructure?

Run "/install clob-microstructure" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Polymarket Clob Microstructure free?

Yes, Polymarket Clob Microstructure is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Polymarket Clob Microstructure support?

Polymarket Clob Microstructure is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Polymarket Clob Microstructure?

It is built and maintained by Mibayy (@mibayy); the current version is v2.0.6.

💬 Comments