The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches and processes news content from 50+ external RSS feeds. Maliciously crafted content in these feeds could attempt to influence the agent's behavior during analysis.
Ingestion points: RSS feed data is fetched in scripts/feed_fetcher.py and parsed in scripts/feed_parser.py.
Boundary markers: The skill does not wrap external news content in specific safety delimiters in its terminal or file output.
Capability inventory: The skill can write processed data to the local file system via the --output argument in scripts/news_aggregator.py and stores temporary data in data/cache.
Sanitization: The skill implements basic sanitization in scripts/feed_parser.py, which removes HTML tags and truncates long summaries to mitigate injection risks.
[EXTERNAL_DOWNLOADS]: The skill connects to established news organizations (such as CoinDesk, The Block, and CoinTelegraph) to retrieve RSS feeds. These downloads target well-known services and are essential for the skill's news aggregation functionality.
[COMMAND_EXECUTION]: The skill executes local Python scripts to perform data aggregation and processing. These operations are restricted to the skill's internal directory and intended functionality.

aggregating-crypto-news