Darwin

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

Annotations already declare destructive and idempotent hints, but the description adds meaningful specifics: every mutation requires requestId, approval uses the exact returned revision and digest, and nonterminal work is never reported as complete. These details are not inferable from annotations and help the agent understand the required protocol and state handling.

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?

The description is two sentences with dense, relevant information, avoiding fluff. The key workflow steps are listed in a compressed but understandable manner. While the first sentence is long, it packs essential details without being verbose.

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?

This is a complex tool with 15 parameters, nested objects, an output schema, and five different operations. The description provides a high-level overview but lacks crucial details like which parameters are required for each type, how to construct inputs per the inputContract, or how the approval digests are obtained. Given the tool's complexity and the zero schema coverage, the description is insufficient for an agent to correctly invoke all operations without additional guidance.

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. It adds meaning for capabilityId and capabilityRevision (from search), and for requestId (required for mutations). However, it does not explain most other parameters such as goalId, inputs, intent, decision, or the approval-related fields, nor how they relate to the different type operations. This is partial compensation for a 15-parameter schema.

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 states the tool makes an outcome happen through a durable Darwin Goal and lists the concrete operations (start, status, respond, approve, stop). This distinguishes it from the sibling search_darwin_capabilities, which is for finding capabilities. However, the phrasing 'Make an outcome happen' is somewhat abstract and the exact resource is only implied by the title and the mention of capabilityId.

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 instructs to start with an exact capabilityId and capabilityRevision from search, clearly indicating that search_darwin_capabilities should be used first. It also enumerates distinct usage flows (poll with status, respond with clarification, approve, stop), providing situational context. It does not explicitly state when not to use this tool, but the search prerequisite covers the main alternative.

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

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

Annotations already declare readOnlyHint, openWorldHint, idempotentHint, and destructiveHint=false. The description adds valuable behavioral guarantees beyond that: results never invent confidence scores, ordering is canonical, and calling with a capabilityId returns inputContract details such as allowed field names, scalar types, constraints, meaning, and required state.

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 sentences, each with distinct value: purpose and scope, execution workflow and return guarantees, and ordering plus anti-hallucination behavior. No filler or redundant restatement of the schema.

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 search tool with 7 optional parameters and nested objects, the description covers the core workflow and critical output behavior. It does not mention pagination or how to use advanced filters, but those are optional and their schema names are reasonably self-explanatory. With an output schema present, the absence of return-structure detail is acceptable.

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 coverage is only 43%, so the description carries a burden it partially meets. It clarifies the roles of query, aiId, and capabilityId, but it does not explain filters, context, cursor, or limit. The schema names hint at their purpose, but the description leaves those to inference rather than compensating for the low coverage.

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 and resource: 'Find executable public Darwin capabilities.' It distinguishes the tool from its sibling by framing it as a pre-execution lookup step that returns exact capability IDs and revisions, and by the contrast with the sibling name execute_darwin_capability.

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?

It gives three concrete invocation modes: natural-language query, aiId to browse one AI, and capabilityId for exact lookup. It also states when to call it relative to execution ('before Act'). It does not explicitly list exclusion cases, but the context is clear enough for selecting this tool over execution.

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