Ask a human to approve or deny an action; returns a URL for them and a signed verdict for you. Times out as denied.
| Name | Required | Description | Default |
|---|---|---|---|
| action | Yes | what you want to do | |
| detail | No | context for the human | |
| webhook | No | called with the verdict | |
| apprise_url | No | ping the human here | |
| timeout_seconds | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full burden. It discloses key behavioral traits: it returns a URL for the human and a signed verdict for the agent, and times out as denied. This goes beyond the schema and gives important expectations, though it could further explain how the verdict is delivered (e.g., via webhook).
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences, front-loaded with the core purpose, and no extraneous words. Every phrase adds value: 'Ask a human', 'returns a URL', 'signed verdict', 'times out as denied'.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description covers the main workflow (request, human action, verdict) and timeout behavior. However, with 5 parameters and no output schema, it leaves gaps such as how the verdict is received (webhook vs direct return) and the role of 'apprise_url' and 'detail'. Adequate but not comprehensive.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 80%, so baseline is 3. The description does not add meaning to individual parameters beyond what the schema provides; it does not clarify the 'action' format or 'timeout_seconds' behavior directly, though 'times out as denied' indirectly relates to timeout.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb ('ask') and resource ('a human') with a clear action ('approve or deny an action'). It also distinguishes itself from siblings like 'notify_human' (which may just notify) and 'check_approval' (which checks status) by focusing on requesting approval.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly implies when to use the tool: when you need human approval for an action. It does not explicitly mention alternatives or exclusions, but the context is clear enough for an agent to recognize this is the approval-request tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Most tools have clearly distinct purposes, but create_wakeup and delayed_delivery both schedule future webhook deliveries, and frontdoor/inbox are similar buffering mechanisms. Overall, descriptions are detailed enough to distinguish most tools.
The majority follow verb_noun snake_case, but there are exceptions like guaranteed_delivery, delayed_delivery, credit_balance, and counter_incr (abbreviation). Still, the overall pattern is readable and predictable.
45 tools is excessive for a single MCP server, likely overwhelming for an agent. While the scope is broad, many tools could be consolidated or split into multiple servers.
The tool set covers a wide range of reliability primitives (queues, checkpoints, counters, watches, approvals) with create/read/manage operations. Gaps like explicit deletes are handled via TTL or generic cancel_resource, so workflows are not dead-ended.