dense_caption
AI agents call dense_caption to retrieve information from Computer Vision MCP Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Dense captioning analyzes and describes image content but does not create, modify, delete, or execute code. It retrieves visual information and generates text output as a side-effect-free read operation. Confidence is slightly reduced because the tool description is empty, but the server context and sibling tools strongly indicate this is a vision analysis tool in the Read category.
From the tool's definition Tool named 'dense_caption' with no description provided. Sibling tools 'caption_image', 'alt_text', and 'image_metadata' all retrieve or analyze image content without modifying data.
Attacks that exploit this kind of access
dense_caption. It is categorised as a Read tool in the Computer Vision MCP Server MCP Server, which means it retrieves data without modifying state.
Register the Computer Vision MCP Server MCP server in PolicyLayer and add a rule for dense_caption: 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 Computer Vision MCP Server. Nothing to install.
dense_caption 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 dense_caption 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 dense_caption. 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.
dense_caption is provided by the Computer Vision MCP Server MCP server (samhains/cv-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 →