promote_to_eval
Take findings from a completed verification cycle and promote them into eval test cases. This is how the inner loop feeds the outer loop: Phase 4 test results become eval cases, Phase 5 checklists become scoring rubrics, Phase 6 edge cases become adversarial eval cases.
This record as markdown: /tools/io-github-homenshum-nodebench/promote-to-eval.md
What promote_to_eval does on Nodebench
AI agents use promote_to_eval to create or update resources in Nodebench, usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Nodebench environment.
Why promote_to_eval is rated Medium
This tool creates or modifies evaluation artifacts (test cases, rubrics, adversarial cases) by promoting data from one phase to another. It is a Write operation because it produces new structured data that can be updated or removed. The severity is medium because misuse could pollute evaluation frameworks or bias test suites, but the effects are contained within the evaluation system and are theoretically reversible.
From the tool's definition The tool 'promote_to_eval' takes findings and 'promote them into eval test cases' and converts test results, checklists, and edge cases into evaluation artifacts.
Attacks that exploit this kind of access
The rule that runs promote_to_eval safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Nodebench, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For promote_to_eval, this is the rule to start with:
promote_to_eval stays usable, but capped: an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Nodebench, apply this rule, and every promote_to_eval call is checked against it from then on.
Questions about promote_to_eval
Take findings from a completed verification cycle and promote them into eval test cases. This is how the inner loop feeds the outer loop: Phase 4 test results become eval cases, Phase 5 checklists become scoring rubrics, Phase 6 edge cases become adversarial eval cases. It is categorised as a Write tool in the Nodebench MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Nodebench MCP server in PolicyLayer and add a rule for promote_to_eval: 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 Nodebench. Nothing to install.
promote_to_eval 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 promote_to_eval 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 promote_to_eval. 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.
promote_to_eval is provided by the Nodebench MCP server (nodebench-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Nodebench, and thousands of servers like it.
This server
Across the catalogue