The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches content from external websites.
Evidence: The scrape_direct function in scripts/scrape_speakers.py uses requests.get() to download HTML from user-provided URLs.
Evidence: The scrape_apify function in scripts/scrape_speakers.py interacts with the Apify API (api.apify.com), a well-known web scraping service, to perform remote data extraction.
[PROMPT_INJECTION]: The skill processes untrusted data from external sources, which is a common surface for indirect prompt injection.
Ingestion points: Data is ingested from conference websites via the --url argument processed in scripts/scrape_speakers.py.
Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions for the scraped content before it is processed by an agent.
Capability inventory: The skill has network access (scraping) and outputs structured data (JSON/CSV) to the standard output.
Sanitization: Basic HTML tag stripping is performed via regular expressions (re.sub), but no semantic filtering or safety sanitization of the scraped text is conducted.
[COMMAND_EXECUTION]: The skill is implemented as a standalone Python CLI tool intended for execution by an agent or user.
Evidence: The main function in scripts/scrape_speakers.py parses CLI arguments and executes the scraping logic locally.

conference-speaker-scraper