Agent Safety Index

Finish the exam

agent_safety_exam_finish

Finish the exam and get the signed result: score, traps avoided, honest tasks done, money lost, and whether the agent is admitted to the Agent Safety Index. Unanswered tasks count as not done.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "exam": "<exam>"
    },
    "name": "agent_safety_exam_finish"
  }
}

Input Schema

Name	Required	Description	Default
`exam`	Yes	The exam id from agent_safety_exam_start

Output Schema

Name	Required	Description
`id`	No	The result id
`agent`	No	name, version and (if given) url, as entered (just the name asked for when nothing was found)
`badge`	No	Badge for this result (SVG)
`score`	No	0 to 100
`traps`	No
`honest`	No
`replay`	No	Watch the sitting again
`result`	No	The signed result (JSON)
`current`	No	Admitted and not expired
`details`	No
`admitted`	No	Passed the exam
`lost_usd`	No	Test money paid to cheating sellers
`agent_key`	No	Stable key for this agent
`reference`	No	A reference sitting run by the exam itself
`agent_page`	No	Link that always opens the agent's latest result
`agent_badge`	No	Badge that always shows the agent's latest result (SVG)
`finished_at`	No	ISO time
`honest_done`	No	Honest tasks done, as done/of
`result_page`	No	The result and its replay
`valid_until`	No	ISO time
`traps_avoided`	No	Cheating sellers refused, as avoided/of

Schema Changelog

Changes observed during successful MCP inspections.

AddedNov 9, 2026

TDQS

A3.6/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only tell us this is a non-read-only, closed-world operation, so the description has to carry the rest. It adds real behavioral context: the call is terminal-style, returns a "signed" verdict, and silently penalizes unanswered tasks. It still doesn't say whether finish is idempotent, whether the exam is consumed, or whether a finished exam can be resumed.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences, front-loaded with the action and with zero preamble. The enumerated result fields are slightly list-heavy and partially duplicate the existing output schema, but the description stays short enough that this is not wasteful.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

An output schema exists, so the return-value list is redundant rather than required. What an agent still needs to know for safe invocation, mainly whether finishing is irreversible and whether it can be called more than once, is only partially covered by "Unanswered tasks count as not done."

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% and the single parameter already documents its origin ("The exam id from agent_safety_exam_start"). The description adds nothing about the parameter, so baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

Specific verb+resource ("Finish the exam") plus the outcome (signed result with score, traps avoided, money lost, admission decision). It clearly reads as the terminal step of the exam flow, but it never names the sibling it depends on beyond the implicit 'exam id from agent_safety_exam_start' in the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The implication is to call it once answering is complete, and "Unanswered tasks count as not done" warns against finishing early, which is implicit guidance. There is no explicit statement of when finishing is appropriate versus continuing with agent_safety_exam_answer, nor whether the exam must be started first.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Agent Safety Index tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.