The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to monitor and process real-time data from external, untrusted sources including Twitter, Reddit, and various news outlets. This creates a surface for indirect prompt injection attacks where malicious instructions hidden in search results or social media posts could attempt to influence the agent's reasoning.
Ingestion points: Data is ingested via tools such as twitter_web_search_timeline, reddit_app_dynamic_search, and google_news_bulk_articles as specified in SKILL.md.
Boundary markers: The instructions lack specific boundary markers or "ignore instructions" wrappers for the fetched content.
Capability inventory: The skill utilizes sandbase_call_tool to execute search queries.
Sanitization: No sanitization or filtering of the retrieved content is mentioned in the provided instructions.
[NO_CODE]: The skill consists exclusively of markdown instruction files (SKILL.md and references/sandbase-api-map.md) and does not include any executable scripts, binaries, or platform-specific configuration files.

event-tracker