frontend-coding-agent

SUSPICIOUS. The core purpose is coherent for a frontend agent, but the skill expands trust by requiring many transitive local skills and by delegating implementation to automated coding agents in full-auto style. No explicit credential harvesting or exfiltration is present, and referenced tooling appears to be official, so this is not malicious; the main risks are transitive trust, autonomous code execution, and prompt-injection exposure from browsing/visual verification workflows.