The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches real-time market activity (merges, splits, and redemptions) from the DOME API at api.domeapi.io. This external communication is the primary function of the skill and follows standard API usage patterns.
[PROMPT_INJECTION]: The skill handles untrusted user-generated content (such as transaction titles and market slugs) from Polymarket, which constitutes an indirect prompt injection surface. The author has implemented proactive defenses to mitigate this risk.
Ingestion points: External data enters the skill via the fetchActivity and fetchAllActivity functions in scripts/activityMonitor.js.
Boundary markers: While explicit prompt delimiters are not provided in the script logic, the data is isolated within structured objects before analysis.
Capability inventory: Analysis of the skill's scripts confirms that its capabilities are limited to data parsing, statistical clustering, and pattern detection; it contains no functions for arbitrary command execution, file system modification, or unauthorized data exfiltration.
Sanitization: The skill utilizes a comprehensive security module (scripts/security.js) that performs regex-based filtering of common injection patterns (e.g., 'ignore previous instructions') and validates the ratio of special characters to identify potential obfuscation attempts.

activity-monitor