Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
polymarket-predictradar-daily-anomalies-skills
v1.0.0Polymarket Daily Anomaly Report. Scans for 3 types of market anomalies: Black Swan (sudden probability shifts in 2h windows), Whale Wars (large opposing bets...
⭐ 0· 38·0 current·0 all-time
byYeri@cnica
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description (daily anomaly report for Polymarket) matches the code and SKILL.md: it queries trade data, classifies domains, and builds reports. However, the skill requires a shared 'polymarket-data-layer' (mcp-client and gamma-client) via relative requires, which implies runtime dependencies outside the skill package. The registry metadata declares no required env vars or credentials, yet the code and usage comments indicate an external MCP_URL and a live MCP service may be required. This is a proportion mismatch between claimed requirements and actual runtime needs.
Instruction Scope
SKILL.md explicitly requires live queries to MCP, Gamma API, and WebSearch and forbids fabricating numbers — consistent with purpose. But it also instructs to display full wallet addresses for copy-trading and to construct clickable Polymarket URLs. The instructions rely on making many SQL queries via mcp.query (including historical queries across wallets) which could surface large volumes of user/market data. The SKILL.md does not declare what credentials or audit/logging policy will be used for WebSearch or MCP access, nor does it limit what context the agent may read beyond the data sources. The instructions reference a shared mcp-client wrapper and handshake behavior (initialize + notifications/initialized), which means the agent must have access to that shared code and any credentials it uses.
Install Mechanism
No install spec — instruction-only plus a single script file. That minimizes disk-install risk. The code uses relative requires (polymarket-data-layer), so risk stems from relying on the host environment to provide that module rather than from an install step downloading arbitrary archives.
Credentials
Registry metadata declares no required env vars or credentials, but the script usage comment ('MCP_URL=... node content-analysis.js') and reliance on mcp-client/gamma-client imply environment configuration and possibly secrets (MCP service URL, API keys) are needed. SKILL.md also expects WebSearch for news context but does not specify which search provider or required API keys. The skill will access wallet addresses and trade histories (sensitive user-related data) but doesn't declare or justify the credentials/privileges needed to query that data—this is an inconsistency and a potential privacy/credential exposure risk.
Persistence & Privilege
The skill does not request always:true and has no install hooks. It appears to be user-invokable and not forced into all agent runs. It does not modify other skills or system-wide settings in the provided files.
What to consider before installing
This skill appears to implement the promised Polymarket anomaly reporting logic, but it depends on a shared 'polymarket-data-layer' (mcp-client and gamma-client) and likely expects environment variables (e.g., MCP_URL and possibly API keys) that are not declared in the skill metadata. Before installing or running: 1) Ask the publisher what runtime dependencies and env vars (MCP_URL, API keys, search provider keys) are required and where the polymarket-data-layer modules come from. 2) Confirm whether mcp-client requires credentials with access to proprietary DBs or cloud services; do not provide high-privilege cloud keys unless you trust the source. 3) Review the full content-analysis.js (the shared module code) and any mcp/gamma clients that will run in your environment to ensure they don't send data to unexpected hosts. 4) Note privacy/ethical implications: the skill displays full 42-char wallet addresses for copy-trading and runs queries across wallet histories (may reveal user-associated activity). If you cannot verify the origin of the shared data-layer or required credentials, run it in an isolated environment or decline installation.Like a lobster shell, security has layers — review code before you run it.
latestvk9736eq5cprxz3evgfg31kmm79842nwj
License
MIT-0
Free to use, modify, and redistribute. No attribution required.