The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, third-party social media platforms (Instagram, Facebook, YouTube, TikTok, etc.) which can contain malicious instructions embedded by an attacker.
Ingestion Point: The script reference/scripts/run_actor.js retrieves data from Apify datasets using the downloadResults and displayQuickAnswer functions.
Capability Inventory: The skill has the capability to execute a Node.js script that performs network requests and writes results to the local filesystem.
Boundary markers: There are no explicit boundary markers or instructions to the agent to disregard instructions found within the downloaded data.
Sanitization: While the script implements protections against spreadsheet formula injection (CSV injection), it does not sanitize the text content for potential prompt injection patterns before the agent processes it.
[COMMAND_EXECUTION]: The skill instructs the agent to execute a local Node.js script with parameters derived from user input.
Evidence: Multiple examples in SKILL.md (Step 4) demonstrate the use of node --env-file=.env to run the analysis script.
The script reference/scripts/run_actor.js implements defensive measures such as regular expression validation for Actor IDs (/^[a-zA-Z0-9_-]+(?:[\/~][a-zA-Z0-9_-]+)?$/) and strict JSON parsing of input to mitigate the risk of command injection.

apify-trend-analysis