discover_capabilitiesSearch the live SqueezeOS capability inventory. Lexical mode is always available; unified mode may add configured evidence-backed semantic/graph context without fabricating scores.
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | lexical | |
| limit | No | ||
| query | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations cover the read-only, idempotent, open-world safety profile, so the description isn't burdened with restating that. It adds a valuable behavioral promise: 'without fabricating scores' – implying the tool returns only evidence-backed results, which is not disclosed elsewhere. This goes beyond the annotations and gives an agent confidence about data integrity.
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 carry zero waste. The purpose is front-loaded, and the mode distinction follows naturally. Each sentence earns its place with concrete information, achieving high information density without bloat.
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 is adequate for a search tool with three obvious parameters, and annotations cover safety. However, it does not mention what the response looks like (e.g., list of capability names, metadata, sorting), which an agent might need to interpret results. Since there is no output schema, this is a gap, but the tool's name and sibling context partially compensate.
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 0%, so the description must illuminate the parameters. It explains the 'mode' parameter well, distinguishing lexical from unified and noting the no-fabrication guarantee. The other parameters (query, limit) are self-explanatory from their names and constraints, so no extra detail is necessary. The explanation of mode fully compensates for its complexity.
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 a specific verb and resource: 'Search the live SqueezeOS capability inventory.' It clearly distinguishes this from sibling tools like inspect_capability (which likely targets a specific capability) and call_capability (which invokes). The mention of two modes further clarifies scope.
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 provides guidance on when to use lexical vs unified mode ('Lexical mode is always available; unified mode may add configured evidence-backed semantic/graph context'), but gives no explicit direction on when to choose this tool over its siblings (e.g., inspect_capability, recommend_capability). The purpose implies search usage, but no exclusions or alternatives are named.
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 clear ownership: capability inventory tools (manifest, discover, inspect, compare, recommend) and x402 payment tools (contract, readiness, evidence) are cleanly separated, with call_capability and agent_web_access distinguished by API-path vs arbitrary-URL scope. Some overlap remains between discover_capabilities, capability_manifest, and recommend_capability, since all three surface capability inventory in slightly different modes.
The capability tools mostly follow a verb_noun pattern (call_capability, compare_capabilities, discover_capabilities, inspect_capability, recommend_capability), but capability_manifest, system_status, agent_web_access, and the x402_* group are noun phrases rather than actions. This is readable and grouped, but not a single predictable convention.
11 tools is well within the ideal range, and each tool maps to a distinct concern: discovery, inspection, recommendation, invocation, status, and x402 payment support. The count feels appropriately scoped for a capability-and-payment-focused server, neither bloated nor too thin.
The tool set covers the full intended workflow: discover, inspect, compare, and recommend capabilities; invoke them with write confirmation and payment-signature handling; and check payment readiness, contracts, and settlement evidence. Management operations are outside the stated purpose because the manifest is explicitly generated from a Flask route table, so there are no obvious dead ends.