The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from TikTok video descriptions and metadata. \n
Ingestion points: External data enters the system through fetch_tiktok.py, which retrieves post content via the Apify TikTok Scraper API. \n
Boundary markers: There are no explicit delimiters or instructions provided to the AI agent to ignore potentially malicious instructions embedded within the video text during analysis or report generation. \n
Capability inventory: The skill instructions include capabilities for file creation (report.md), local script execution, and interaction with an external AI analysis service (video-content-analyzer). \n
Sanitization: While analyze_posts.py uses regex for keyword extraction, there is no sanitization or escaping of the raw video text before it is passed to the AI for structure and hook analysis. \n- [EXTERNAL_DOWNLOADS]: The skill fetches video metadata and engagement metrics from Apify's cloud-based scraping infrastructure. This is an expected behavior for a research tool leveraging well-known third-party data providers. \n- [COMMAND_EXECUTION]: The defined workflow requires the execution of Python scripts provided within the skill's package. These scripts manage data retrieval and statistical analysis within localized run folders.

tiktok-research