The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands. This includes using the mcpc CLI tool to fetch actor schemas and running a local Node.js script (run_actor.js) to trigger data extraction tasks.
[DATA_EXFILTRATION]: The skill is designed to read the APIFY_TOKEN from a local .env file and transmit it to api.apify.com to authenticate requests. This is a functional requirement for interacting with the service.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @apify/mcpc package from the npm registry, which is the official tool for managing Apify Model Context Protocol connections.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it retrieves and processes data from various external platforms (Google Maps, Facebook, Instagram, TikTok, etc.) via Apify actors.
Ingestion points: Untrusted data enters the agent's context through the script's output in Step 4 of the workflow (displayed in chat or read from generated files).
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are present in the prompts that handle the scraped data.
Capability inventory: The skill possesses the capability to write to the file system (via writeFileSync in run_actor.js) and execute shell commands (as defined in SKILL.md).
Sanitization: While the script formats data into JSON or CSV, it does not sanitize the content of the scraped fields to prevent the agent from interpreting embedded text as instructions.

apify-competitor-intelligence