finding-fitness-brands-on-tiktok

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from TikTok, such as video descriptions and channel bios, which are untrusted external sources.
  • Ingestion points: Data returned from the TikTok Scraper actor (SKILL.md).
  • Boundary markers: None identified to separate instructions from processed data.
  • Capability inventory: The skill uses subprocess execution (node scripts), file system writing (CSV/JSON output), and network operations (curl/Apify API).
  • Sanitization: No specific sanitization or filtering logic is described for the ingested TikTok content.
  • [DYNAMIC_EXECUTION]: The customMapFunction input parameter allows the agent to pass a JavaScript function as a string to be executed on the Apify platform. While this is a feature of the underlying tool, providing an interface for raw code injection into a remote execution environment is a dynamic execution surface.
  • [COMMAND_EXECUTION]: The documentation instructs the agent to execute shell commands using node scripts/run_actor.js and curl. These commands interact with the local file system (saving results) and the network (communicating with Apify's API).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 07:59 PM
Security Audit — agent-trust-hub — finding-fitness-brands-on-tiktok