Adversarial behavioural-bias engine — audits your decisions for cognitive biases via your own AI.
Scored across 3 tools
Each tool has a distinct purpose: analyzing decisions, retrieving a bias by ID, and listing biases. There is no overlap or ambiguity.
All tools follow a consistent verb_noun pattern in snake_case (analyze_decision, get_bias, list_biases), making the naming predictable and clear.
With 3 tools, the server is well-scoped for its domain of bias analysis. The count is neither too sparse nor excessive.
The tool surface covers the core operations: listing biases, retrieving details, and analyzing decisions. Minor gaps exist (e.g., no tool to search or filter more granularly), but the set is sufficient for the intended use.
Adversarially audit a decision for cognitive biases. Returns a directive YOUR model executes to produce the composed audit (verdict-first). Provide reasoning, not just the conclusion.
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | forward = a decision you are about to make; retrospective = reviewing a past decision/outcome. | |
| judgment | Yes | The decision/judgment in one line. | |
| language | No | Optional. Natural language for the audit prose (e.g. "Tamil", "Spanish"). Bias ids stay canonical English so the result is still parseable. Defaults to English. | |
| reasoning | No | How you arrived at it (required to audit). | |
| structured | No | Optional. Default false → the audit comes back as readable prose. Set true to get a machine-parseable JSON object (bias ids/keys in English) for pipelines that store or compare audits. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations exist; description reveals it returns a directive the model executes, which is a key behavioral trait. However, it does not disclose side effects, safety, or whether it modifies 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?
Two sentences, front-loaded with purpose, no redundancy. Every sentence adds value.
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?
Adequately describes purpose and output nature ('directive', 'verdict-first'), but lacks details on output structure and expected behavior in edge cases. No output schema to 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 covers all parameters (100%). Description adds value on 'reasoning' parameter ('Provide reasoning, not just the conclusion') but does not enrich other parameters beyond 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?
Clearly states it adversarially audits a decision for cognitive biases using specific verb 'audit' and resource 'decision'. Distinguishes from sibling tools get_bias/list_biases which deal with bias definitions.
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?
Implies use when auditing a decision, but no explicit when-to-use, when-not-to-use, or alternatives provided. The instruction 'Provide reasoning, not just the conclusion' gives partial guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get the full entry for one bias by id.
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must fully convey behavioral traits. It mentions a read operation ('get'), but does not disclose any additional details such as required permissions, data size limits, or response characteristics. This lack of transparency is a significant gap for a tool with no 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?
The description is a single, concise sentence that efficiently conveys the tool's purpose and key input. Every word earns its place with no redundancy or unnecessary detail.
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?
Given the tool has only one parameter and no output schema, the description provides a minimal but acceptable level of completeness. It explains what the tool does and what input it requires, but lacks information about the output format or any constraints. This is adequate for a simple lookup tool but could be more informative.
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?
With schema description coverage at 0% and only one parameter 'id' (type: string), the description adds minimal semantic value by specifying 'by id'. However, it does not clarify the format or source of the id, nor how to obtain it. For a single-parameter tool, this is adequate but not exceptional.
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 'Get the full entry for one bias by id' clearly specifies the action (get), resource (full entry for one bias), and required input (id). It effectively distinguishes from siblings like list_biases (which likely lists all) and analyze_decision (which analyzes a decision).
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 explicitly states the scenario for using this tool: retrieving a single bias's full details by its id. While it does not provide explicit exclusion criteria or alternatives, the context of sibling tools list_biases and analyze_decision implicitly guides usage for single-object retrieval.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List the 22-bias catalogue, optionally filtered by family.
| Name | Required | Description | Default |
|---|---|---|---|
| family | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are present, and the description only states the tool lists biases. It does not disclose order, pagination, limits, or any side effects. The description carries the full burden for behavioral transparency but falls short.
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 concise sentence with no filler. It quickly conveys the core function and the only parameter.
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?
No output schema is provided, and the description does not specify what fields or structure the list returns. For a listing tool, this is a significant gap, as the agent needs to know what to expect.
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 mentions the optional family filter, but does not explain the meaning of each family enum value beyond the enum labels themselves, which are self-explanatory. Minimal added value.
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 lists the 22-bias catalogue with optional filtering by family. This distinguishes it from siblings 'get_bias' (single bias retrieval) and 'analyze_decision' (analysis function).
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 explicit when-to-use or alternative guidance is provided. However, the description implies it is for browsing the full catalogue, while 'get_bias' likely targets a specific bias. This is intuitive but not stated.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
analyze_decisionget_biaslist_biasesTrust gate for AI agents: multi-model adversarial consensus, signed and verifiable verdicts.
Personalised AI augmentation system — makes you better at your work, not dependent on AI
Convene a panel of expert AI personas to debate any decision from every side.
The decision layer for AI agents: act, escalate or refuse, and every decision comes back signed.
Add one secure layer between your agents and this server.