{
  "$schema": "https://a2a-protocol.org/schemas/v0.3/agent-card.json",
  "protocolVersion": "0.3",
  "id": "styxx",
  "name": "styxx",
  "description": "Cognitive vitals for language models. Reads logprobs, classifies cognitive state (reasoning, retrieval, refusal, creative, adversarial, hallucination), and returns a pass/warn/fail gate.",
  "version": "3.4.0",
  "tagline": "nothing crosses unseen.",
  "documentationUrl": "https://fathom.darkflobi.com/styxx",
  "url": "https://fathom.darkflobi.com/styxx",
  "provider": {
    "id": "fathom-lab",
    "name": "Fathom Lab",
    "url": "https://fathom.darkflobi.com"
  },
  "defaultInputModes": ["application/json"],
  "defaultOutputModes": ["application/json"],
  "capabilities": {
    "streaming": false,
    "pushNotifications": false,
    "stateTransitionHistory": false
  },
  "securitySchemes": {},
  "security": [],
  "skills": [
    {
      "id": "observe",
      "name": "Observe cognitive vitals",
      "description": "Reads logprobs from a model response and returns Vitals: classification, confidence (0-1), and a pass|warn|fail gate.",
      "tags": ["vitals", "classification", "logprobs", "interpretability"],
      "examples": [
        "Observe this response and tell me if it looks like hallucination.",
        "What is the cognitive state of this completion?"
      ],
      "inputModes": ["application/json"],
      "outputModes": ["application/json"],
      "input": {
        "type": "object",
        "additionalProperties": false,
        "required": ["response"],
        "properties": {
          "response": {
            "type": "object",
            "description": "OpenAI-compatible chat-completion response containing logprobs."
          }
        }
      },
      "output": { "$ref": "#/components/schemas/Vitals" }
    },
    {
      "id": "verify",
      "name": "Verify response trustworthiness",
      "description": "Runs trajectory analysis over a response's logprob stream and returns a VerificationResult with confidence and anomaly flags.",
      "tags": ["verification", "trajectory", "confidence"],
      "examples": [
        "Verify that this answer is trustworthy.",
        "Does this completion show confabulation patterns?"
      ],
      "inputModes": ["application/json"],
      "outputModes": ["application/json"],
      "input": {
        "type": "object",
        "additionalProperties": false,
        "required": ["response"],
        "properties": {
          "response": {
            "type": "object",
            "description": "OpenAI-compatible chat-completion response with logprobs."
          }
        }
      },
      "output": { "$ref": "#/components/schemas/VerificationResult" }
    },
    {
      "id": "classify",
      "name": "Classify trajectory shape",
      "description": "Given a sequence of token logprobs, returns the cognitive class (reasoning, retrieval, refusal, creative, adversarial, hallucination).",
      "tags": ["classification", "trajectory"],
      "inputModes": ["application/json"],
      "outputModes": ["application/json"],
      "input": {
        "type": "object",
        "additionalProperties": false,
        "required": ["logprobs"],
        "properties": {
          "logprobs": {
            "type": "array",
            "items": { "type": "number" },
            "minItems": 1
          }
        }
      },
      "output": {
        "type": "object",
        "additionalProperties": false,
        "required": ["classification", "confidence"],
        "properties": {
          "classification": {
            "type": "string",
            "enum": ["reasoning", "retrieval", "refusal", "creative", "adversarial", "hallucination"]
          },
          "confidence": { "type": "number", "minimum": 0, "maximum": 1 }
        }
      }
    },
    {
      "id": "weather",
      "name": "Cognitive weather report",
      "description": "Returns a fleet-level weather report summarising recent cognitive vitals.",
      "tags": ["monitoring", "fleet"],
      "inputModes": ["application/json"],
      "outputModes": ["application/json"],
      "input": {
        "type": "object",
        "additionalProperties": false,
        "properties": {
          "window": { "type": "integer", "minimum": 1, "default": 100 }
        }
      },
      "output": {
        "type": "object",
        "additionalProperties": true,
        "properties": {
          "summary": { "type": "string" },
          "gate": { "type": "string", "enum": ["pass", "warn", "fail"] }
        }
      }
    }
  ],
  "components": {
    "schemas": {
      "Vitals": {
        "type": "object",
        "additionalProperties": false,
        "required": ["classification", "confidence", "gate"],
        "properties": {
          "classification": {
            "type": "string",
            "enum": ["reasoning", "retrieval", "refusal", "creative", "adversarial", "hallucination"]
          },
          "confidence": { "type": "number", "minimum": 0, "maximum": 1 },
          "gate": { "type": "string", "enum": ["pass", "warn", "fail"] }
        }
      },
      "VerificationResult": {
        "type": "object",
        "additionalProperties": false,
        "required": ["valid", "confidence"],
        "properties": {
          "valid": { "type": "boolean" },
          "confidence": { "type": "number", "minimum": 0, "maximum": 1 },
          "gate": { "type": "string", "enum": ["pass", "warn", "fail"] },
          "classification": { "type": "string" },
          "anomalies": { "type": "array", "items": { "type": "string" } },
          "trajectory": {
            "type": "object",
            "additionalProperties": true,
            "properties": {
              "slope": { "type": "number" },
              "curvature": { "type": "number" },
              "volatility": { "type": "number" }
            }
          }
        }
      }
    }
  }
}