prediction-markets-analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests and analyzes public Kalshi market pages (see SKILL.md example queries such as "Analyze this... https://kalshi.com/...") and documents tools (e.g., octagon-deep-research-agent and prediction_markets_history) that aggregate web/public sources, so untrusted third-party content is read and can materially influence recommendations and actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill configures and invokes Octagon MCP at runtime (e.g., the command "npx -y octagon-mcp" which pulls/executes the octagon-mcp package and references the project at https://github.com/OctagonAI/octagon-mcp), and it also supports connecting to the remote MCP endpoints (https://mcp.octagonagents.com/mcp or https://mcp.octagonagents.com/{OCTAGON_API_KEY}/mcp), meaning remote code/behavior is fetched or executed at runtime and can directly influence agent instructions.