memory_getFetch one stored value by its exact key.
When to use: Call when you know the key you wrote earlier. Cheaper and more precise than memory.search.
Price: US$0.000500 per call.
| Name | Required | Description | Default |
|---|---|---|---|
| key | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true and openWorldHint=false, so the safety profile is covered. The description adds genuine behavioral context beyond that: exact match only (no partial lookup), the requirement that the key must have been previously written, and an explicit price per call. No contradiction with annotations.
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?
Three short sections with zero filler: purpose first, then usage guidance, then cost. Every sentence earns its place and the most important scoping claim is front-loaded.
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?
For a one-parameter read-only tool, the description covers purpose, when to use, the correct sibling alternative, and cost. Return format remains unspecified, but readOnlyHint plus the simple retrieval semantics make the call safe and predictable without it.
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 0%, so the description must compensate, and it does: 'exact key' adds match semantics and 'the key you wrote earlier' ties the parameter to a prior memory_set call. The schema only provides type and length constraints, so the description meaningfully enriches the meaning of the single parameter.
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 opens with a specific verb+resource statement: 'Fetch one stored value by its exact key,' which clearly scopes the operation to exact-key retrieval of a single value. The title 'Recall by key' reinforces this, and the usage line explicitly contrasts it with memory.search so an agent can distinguish the tools.
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 'When to use' line gives an explicit condition ('when you know the key you wrote earlier') and names the alternative (memory.search) with a reason to prefer this tool ('Cheaper and more precise'). This tells the agent not just when, but why to select it over a sibling.
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.
Each tool addresses a distinct resource or action: memory operations are clearly separated from inbox operations, await/park/resume are differentiated by whether the process stays alive or hands off state, and administrative calls like capabilities, whoami, and credit_deposit have non-overlapping purposes. No two tools appear interchangeable.
Most resource operations follow an object_verb pattern (memory_get, hook_create, inbox_list), but core primitives like await, park, resume, and signal are bare verbs, while capabilities and whoami are bare nouns. The style is consistently lowercase snake_case and readable, but the naming convention is mixed rather than uniform.
At 16 tools the set is just past the typical well-scoped range, but each tool maps to a distinct platform capability: identity, inbox, memory, hooks/signals, parking, and billing. It feels slightly heavy rather than bloated.
The core agent workflow is covered end to end: identity creation, credit deposit, hook creation, awaiting triggers, inbox reading, durable memory, and the park/resume round-trip. Minor gaps remain—there is no hook list/delete and no way to cancel a parked state—but they can be worked around rather than blocking.