The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the AI agent to execute shell commands using node and curl. These commands are used to interface with a local script (scripts/run_actor.js) and the Apify REST API to perform data scraping tasks.
[DYNAMIC_EXECUTION]: The input parameter customMapFunction accepts a raw string representing a JavaScript function. This string is passed to the remote execution environment (the Apify actor) where it is executed to transform data. While this is a standard feature of the underlying platform, it constitutes a dynamic code execution surface where unvalidated user input could be processed.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user, specifically searchTerms, usernames, and topic keywords. These values are interpolated into shell commands and API payloads. While the examples use quotes and JSON structures to mitigate risk, this remains a surface for potential injection if delimiters are bypassed.
Ingestion points: Topic keywords, industry names, and Twitter handles provided by the user in Step 1 and Step 2.
Boundary markers: The skill uses JSON objects and single quotes in shell commands to wrap user input.
Capability inventory: Uses node to execute local scripts and curl for network operations.
Sanitization: No explicit sanitization logic is described beyond the use of standard formatting.

tracking-twitter-thought-leaders