AI agents call html_decode to retrieve information from Devutils without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs a pure in-memory transformation of input text—converting HTML entities to characters—with no side effects, no data storage, no external calls, and no destructive operations. It is functionally equivalent to a read/transform operation on local data.
From the tool's definition "Decode HTML entities in a string back to their original characters"
Attacks that exploit this kind of access
Decode HTML entities in a string back to their original characters. It is categorised as a Read tool in the Devutils MCP Server, which means it retrieves data without modifying state.
Register the Devutils MCP server in PolicyLayer and add a rule for html_decode: 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 Devutils. Nothing to install.
html_decode 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 html_decode 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 html_decode. 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.
html_decode is provided by the Devutils MCP server (paladini/devutils-mcp-server). 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 →