finding-brand-ambassador-candidates

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Instagram and TikTok posts, including captions and metadata, which could contain adversarial instructions targeting the agent.
  • Ingestion points: Scraped content from Instagram and TikTok via Apify actors.
  • Boundary markers: None identified in the workflow or output instructions to separate untrusted data from agent instructions.
  • Capability inventory: Execution of local scripts via node, network requests via curl, and use of the apify:run-actor tool.
  • Sanitization: No evidence of sanitization, escaping, or filtering of the scraped content before processing or presentation.
  • [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands to interact with external services.
  • Evidence: Shell command blocks in SKILL.md using node scripts/run_actor.js and curl for REST API fallback to the Apify platform.
  • [EXTERNAL_DOWNLOADS]: The skill interacts with external actors to perform data scraping.
  • Evidence: References to apidojo~instagram-scraper and apidojo~tiktok-scraper on the Apify platform. These are vendor-owned resources from 'apidojo-io' used for their primary purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 06:56 PM
Security Audit — agent-trust-hub — finding-brand-ambassador-candidates