Check if input contains an attack (without bending). Returns attack type if found.
AI agents call check_input to retrieve information from Inject Bender without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only reads and analyzes input to detect attack patterns, then returns information about those patterns. It has no capability to execute code, modify data, delete records, or initiate external operations. The lack of side effects and read-only nature place it firmly in the Read category with low severity.
From the tool's definition Tool performs detection/checking of input for attacks and 'returns attack type if found' — a pure informational query with no side effects.
Attacks that exploit this kind of access
Check if input contains an attack (without bending). Returns attack type if found. It is categorised as a Read tool in the Inject Bender MCP Server, which means it retrieves data without modifying state.
Register the Inject Bender MCP server in PolicyLayer and add a rule for check_input: 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 Inject Bender. Nothing to install.
check_input 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 check_input 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 check_input. 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.
check_input is provided by the Inject Bender MCP server (jaspertvdm/mcp-inject-bender). 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 →