AI agents call get_keywords to retrieve information from YT-NINJA without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs keyword extraction, which is purely a data retrieval and analysis operation. It does not modify video data, execute commands, delete content, or trigger financial transactions. The blast radius of misuse is minimal — an attacker could extract keywords from videos without authorization, but this would only constitute unauthorized information access, not data destruction or system compromise.
From the tool's definition Tool name is 'get_keywords' and description states 'Extract keywords from video' — a read-only operation that retrieves derived data (keywords) from existing video content without modification, deletion, or external execution.
Attacks that exploit this kind of access
Extract keywords from video. It is categorised as a Read tool in the YT-NINJA MCP Server, which means it retrieves data without modifying state.
Register the YT-NINJA MCP server in PolicyLayer and add a rule for get_keywords: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches YT-NINJA. Nothing to install.
get_keywords is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the get_keywords rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for get_keywords. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
get_keywords is provided by the YT-NINJA MCP server (satyamkumar420/yt-ninja-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →