alterlab-perplexity

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly performs real-time web searches and ingests web-grounded answers from Perplexity models via OpenRouter (stated in SKILL.md and implemented in scripts/perplexity_search.py which calls litellm completion with model "openrouter/perplexity/*"), meaning it fetches untrusted public web content that the agent reads and uses to produce answers and drive workflows—exposing it to indirect prompt injection.