playgama
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a standard implementation of an MCP (Model Context Protocol) server designed to interface with playgama.com. All tools and remote endpoints reside on
unbrowse.ai, which is the official domain of the skill author. - [CREDENTIALS_UNSAFE]: The instructions explicitly promote secure credential management, advising users to store API keys in secret managers rather than configuration files. It also includes a strict prohibition against the agent asking users for passwords, utilizing a vendor-managed autofill system instead.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external content from playgama.com via the
playgama_com__read_pageandplaygama_com__get_searchtools. While this data is untrusted, the skill incorporates mitigation strategies by requiring human authorization for any sensitive or state-changing actions (such as posting or purchasing) and instructing the agent to stay within the domain scope. - [COMMAND_EXECUTION]: The provided installation and configuration commands (e.g.,
claude mcp add) are standard administrative procedures for setting up MCP servers and do not involve arbitrary or malicious shell execution.
Audit Metadata