apify-trend-analysis
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes data fetched from social media platforms (TikTok, Instagram, YouTube, etc.) which are untrusted sources. This data is presented directly to the agent, creating a surface for indirect prompt injection.\n
- Ingestion points: Data returned from the Apify API and displayed in the agent's context or saved to local files via run_actor.js.\n
- Boundary markers: The output script uses visual separators (===) but does not include explicit instructions for the agent to treat the fetched content as data rather than instructions.\n
- Capability inventory: The skill allows the agent to execute shell commands (Node.js scripts and the mcpc CLI) and perform file system writes.\n
- Sanitization: The script truncates long strings and implements CSV injection protection, but it does not sanitize fetched content for potential prompt injection patterns.\n- [EXTERNAL_DOWNLOADS]: The skill fetches Actor schemas and trend data from Apify's official services (api.apify.com and mcp.apify.com). It also utilizes the @apify/mcpc CLI tool from the official NPM registry.\n- [COMMAND_EXECUTION]: The skill workflow involves executing shell commands, including the mcpc CLI and a local Node.js script. While the script itself uses standard argument parsing, the command templates provided in the instructions use shell variables and quotes that require the agent to correctly manage parameter substitution to avoid potential argument injection.
Audit Metadata