agent-browser

The skill aligns with its stated purpose of browser automation for AI agents and includes reasonable safeguards (domain allowlists, action policies). The primary security concerns are typical for developer tooling (supply-chain risk from external repository, need for provenance/verification, and ensuring minimal data exposure of private pages). Overall, the footprint is coherent with its purpose, but maintainers should ensure verifiable integrity of the downloaded CLI (e.g., checksums, signatures) and enforce strict environment controls to prevent unintended data leakage from private sites.