Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed Polymarket data and local paper-trading helper, with no evidence of real trading, credential use, exfiltration, or hidden automatic persistence.
Install only if you are comfortable with a tool that queries public Polymarket data and stores your watched markets plus simulated portfolio history in ~/.polymarket/. It does not place real trades or use credentials, but the paper-trading and cron examples should be treated as user-directed features rather than passive read-only lookup.
The manifest description frames the skill as read-only market research, but the documented commands also create persistent watchlists, store portfolio history, and support simulated buy/sell operations. This mismatch can mislead users or orchestrators into invoking a skill with side effects they did not expect, which is especially risky when local persistence and scheduled execution are involved.
Referenced artifact was not completely inspected
Review `scripts/polymarket.py` before first use to verify behavior.
The manifest describes monitoring and analysis of Polymarket data, but the code also implements buy/sell style paper-trading and portfolio management workflows. Even though no real trades are executed, these features materially change the skill's scope into simulated financial activity, which can mislead users, trigger unintended financial decision support use, and conceal higher-risk functionality behind an underspecified description.
The skill declares no explicit tool scope even though its documented behavior includes network access and local file reads/writes. In an agent ecosystem, missing permission boundaries can allow broader-than-expected execution and reduce the user's ability to assess or constrain what the skill may do.
The manifest description is broad, noisy, and lacks clear invocation constraints, which makes it harder for an agent or user to determine when the skill should be used. Overbroad descriptions increase the chance of inappropriate invocation and can mask capabilities that go beyond simple data retrieval.
The skill metadata presents this as a market data access and analysis tool, but the implementation also creates a persistent local state directory and stores watchlist and portfolio JSON files under the user's home directory. This expands the skill's behavior beyond passive data retrieval into local data retention, which can surprise users, create privacy/forensics concerns, and violate least-astonishment expectations for an analysis-only skill.
No suspicious patterns detected.