The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute a local Node.js script (run_actor.js) and the @apify/mcpc CLI tool. These tools are used to orchestrate data extraction tasks on the Apify platform.
[EXTERNAL_DOWNLOADS]: The run_actor.js script fetches data and actor configurations from api.apify.com and mcp.apify.com. These are official endpoints for Apify, a well-known web scraping and automation platform.
[DATA_EXFILTRATION]: The skill transmits the user's APIFY_TOKEN to the official Apify API to authenticate requests. This is a standard and necessary operation for the skill's functionality. The instructions correctly advise users to store this sensitive token in a .env file.
[PROMPT_INJECTION]: The skill processes data from social media platforms which acts as a surface for indirect prompt injection.
Ingestion points: External social data is ingested via the Apify API in run_actor.js and displayed to the agent or saved to files.
Boundary markers: No explicit delimiters or boundary markers are used in the prompt instructions to isolate potentially malicious external data.
Capability inventory: The run_actor.js script has the capability to write extraction results directly to the local filesystem using writeFileSync.
Sanitization: The script implements csvCell to mitigate CSV formula injection and truncates long strings to reduce the impact of large or malicious inputs.

apify-audience-analysis