The Agent Skills Directory

COMMAND_EXECUTION (MEDIUM): The skill constructs shell commands by interpolating variables like ACTOR_ID and JSON_INPUT directly into the command string. If these variables are derived from untrusted user input or external data, it could lead to command injection.
CREDENTIALS_UNSAFE (MEDIUM): The workflow explicitly requires reading an APIFY_TOKEN from a .env file and passing it through a shell pipeline (grep | xargs). While required for functionality, this pattern exposes secrets to the shell environment and process history.
EXTERNAL_DOWNLOADS (LOW): The skill instructs users to install an external package @apify/mcpc globally via npm. While Apify is a known service, this package is not from the predefined list of trusted organizations.
INDIRECT_PROMPT_INJECTION (LOW): The skill processes data fetched from external social media scrapers (Instagram, TikTok, etc.). This data is inherently untrusted.
Ingestion points: Data returned from mcpc and run_actor.js outputs.
Boundary markers: None identified in the provided instructions.
Capability inventory: Subprocess execution via node and mcpc CLI.
Sanitization: No explicit sanitization or validation of the fetched data is mentioned before processing or summarization.

apify-influencer-discovery