Refresh the saved evaluation for a single paragraph while still using the full chapter as context, including objective scores, writing-style-aware editorial notes, canon coherence checks, a weighted verdict, an explanation of why that verdict landed there, and concrete next steps.
AI agents use evaluate_paragraph to create or update resources in Narrarium — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Narrarium environment.
The tool reads paragraph and chapter content to perform analysis, but its primary side effect is writing/updating a saved evaluation object in the repository. This makes it a Write operation rather than pure Read. Misuse risk is moderate since it overwrites editorial metadata but does not delete content or have financial implications.
From the tool's definition 'Refresh the saved evaluation for a single paragraph' — the tool updates/overwrites a stored evaluation record
Attacks that exploit this kind of access
Refresh the saved evaluation for a single paragraph while still using the full chapter as context, including objective scores, writing-style-aware editorial notes, canon coherence checks, a weighted verdict, an explanation of why that verdict landed there, and concrete next steps. It is categorised as a Write tool in the Narrarium MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Narrarium MCP server in PolicyLayer and add a rule for evaluate_paragraph: 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 Narrarium. Nothing to install.
evaluate_paragraph is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the evaluate_paragraph 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 evaluate_paragraph. 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.
evaluate_paragraph is provided by the Narrarium MCP server (narrarium-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 →