cheminformatics

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (medium risk: 0.65). Both links warrant caution: googleadsagent.ai looks like possible brand-impersonation/typosquatting (higher risk of social‑engineering or malicious downloads), and the GitHub repo is on a reputable platform but is from an unfamiliar account—GitHub code is lower risk overall but unknown repos or their releases/binaries should be inspected carefully before downloading or executing.

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md explicitly states it "integrates with databases like PubChem and ChEMBL for compound retrieval and benchmarking" (Description/Implementation), which are public third‑party repositories whose user-submitted content the agent would fetch and interpret as part of its workflow and could materially influence screening and decision-making.