The Agent Skills Directory

[PROMPT_INJECTION]: The skill retrieves and processes untrusted third-party content from social media platforms and news aggregators, presenting a surface for indirect prompt injection.
Ingestion points: External data enters the agent context via multiple tools, including lunar_topic_posts, lunar_topic_news, lunar_category_posts, and lunar_search_content, which fetch raw text from the LunarCrush API (documented in tools/topics.py, tools/creators.py, and tools/coins.py).
Boundary markers: The skill does not implement explicit delimiters or instructions to the model to ignore potential commands embedded in the retrieved social content.
Capability inventory: Across all files, the skill is limited to performing data retrieval and formatting. There are no tools for file system modification, arbitrary command execution, or other high-risk capabilities that could be exploited via injection.
Sanitization: Retrieved data is formatted into structured dictionaries but does not undergo sanitization to filter out malicious natural language instructions.

@554/lunarcrush