The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill requires the manual extraction and use of the auth_token cookie from the user's browser DevTools to authenticate scraping sessions. This involves the direct handling of sensitive session credentials by the agent or user.
[EXTERNAL_DOWNLOADS]: The instructions specify installing the xactions package from the NPM registry. This is an external third-party dependency not originating from a pre-verified trusted source.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest large amounts of untrusted data from X/Twitter, including tweets, profiles, and direct messages, which could contain malicious instructions designed to influence the agent.
Ingestion points: External data is ingested from x.com timelines, search results, bookmarks, and notifications as described in SKILL.md and references/scraper-details.md.
Boundary markers: There are no identified delimiters or warnings to the agent to ignore instructions embedded within the scraped content.
Capability inventory: The skill possesses file-writing capabilities (exportToJSON, exportToCSV) and network scraping capabilities via Puppeteer.
Sanitization: The analysis found no evidence of sanitization, filtering, or escaping of the scraped content before it is processed or exported.
[DYNAMIC_EXECUTION]: The skill provides various browser console scripts (IIFEs) in the scripts/ directory that are intended to be pasted and executed directly within the browser's developer console. This bypasses standard execution environments and encourages the execution of code snippets outside of formal package management.

twitter-scraping