Skillv2.0.0

ClawScan security

Agent Casino · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousFeb 25, 2026, 11:54 AM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill's instructions match a crypto gambling service (so behavior is coherent) but the source/homepage are unknown and it directs the agent to register, store an API key, and deposit real cryptocurrency — which carries financial risk and warrants extra caution.
Guidance: This skill is coherent with a crypto gambling service, but exercise caution before installing or using it: 1) Verify the external service (agent.rollhub.com) independently — the registry entry lacks a homepage/source so confirm it's legitimate (look for TLS certificate, reputation, official docs). 2) Never give your primary wallet/API keys to an agent you do not fully trust — use a throwaway wallet with minimal funds for testing. 3) Understand the irreversible financial risks: depositing crypto sends real money to an external address. 4) Check legal/regulatory implications in your jurisdiction. 5) If you must use it, restrict permissions, rotate keys, and avoid allowing autonomous agents to manage funds without explicit, narrow authorization. 6) Verify the 'provably fair' proofs returned by the service before relying on them, and be wary of affiliate/referral incentives that may encourage broad distribution.

Review Dimensions

Purpose & Capability: noteThe name/description and SKILL.md consistently describe a provably-fair crypto gambling API (register, deposit, bet, verify, withdraw). Required resources (none declared) are consistent with an instruction-only integration. However, the registry metadata shows no homepage or source, so the external service's provenance is not verifiable from this package.
Instruction Scope: noteAll runtime instructions are network calls to https://agent.rollhub.com and concern registration, depositing crypto to an address, placing bets, verifying bets, and withdrawals. The instructions do not request reading local files or unrelated env vars. Because these steps involve irreversible monetary transfers and saving an API key, they have higher real-world risk even though they stay within the stated gambling scope.
Install Mechanism: okNo install spec and no code files (instruction-only). This minimizes on-disk execution risk; the skill's actions are limited to HTTP calls described in SKILL.md.
Credentials: okThe skill declares no required environment variables or credentials. The SKILL.md instructs obtaining and storing an X-API-Key returned by the service, which is expected for a remote API. There are no unexplained requests for unrelated credentials or system config.
Persistence & Privilege: notealways:false (no forced presence). Model invocation is permitted (platform default). If an agent were given the service API key, it could autonomously place bets and move funds — a significant capability. This is a property of the platform rather than the skill, but combined with the financial nature of this skill it should be treated carefully.