Detect quality drift between current and previous LLM responses
| Name | Required | Description | Default |
|---|---|---|---|
| threshold | No | Drift threshold (0-1, default: 0.15) | |
| currentResponse | Yes | Current LLM response | |
| previousResponse | Yes | Previous LLM response |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It only states a generic action ('detect') without explaining the output format, threshold semantics, side effects, or how drift is measured, leaving significant behavioral ambiguity.
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 a single, front-loaded sentence with no filler or redundant information. It conveys the core purpose efficiently, making it fully concise.
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 tool has no output schema and no annotations, so the description should explain return values or operational context. It does not, leaving the agent uncertain about what the tool returns or how to interpret the detection result. This gap makes the description contextually incomplete for a tool with this level of structured context.
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?
The input schema already provides complete descriptions for all three parameters with 100% coverage, earning a baseline score of 3. The tool description adds no extra parameter semantics beyond what the schema already includes, though it does implicitly reference current and previous responses.
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 clearly states the tool's purpose with a specific verb ('Detect') and resource ('quality drift between current and previous LLM responses'), making it distinct from the sibling tools. However, it does not explicitly differentiate itself from consistency_check or quality_validator, so it misses the top score.
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?
No usage guidance is provided. The description does not mention when to use this tool versus alternatives like consistency_check or quality_validator, nor any prerequisites or conditions for use.
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.
The tools are mostly distinct, but consistency_check and drift_detector both involve comparing responses, and quality_validator and schema_enforcer overlap on validation. The descriptions help clarify boundaries, especially for hallucination_scorer which is clearly unique.
Most names follow a noun_noun pattern (drift_detector, hallucination_scorer, quality_validator, schema_enforcer), but consistency_check deviates by using a verb as the second element. All are snake_case, so the overall style is recognizable but not perfectly uniform.
Five tools is well within the optimal range for a focused monitoring server. Each tool addresses a distinct quality aspect without unnecessary bloat or sparsity.
The toolset covers key monitoring dimensions: single-response quality, schema validation, hallucination risk, cross-response consistency, and time-based drift. Missing semantic hallucination detection is acknowledged as a limitation, but it's a minor gap given the scope.