Artificial eXperience Intelligence (AXI)

AXI design reviews, gated AXI Score, and paper search for AI assistants, agents, and robots.

Sollte ich dies verwenden

Qualität und Sicherheit

A
Qualität der Beschreibung
92%
Vollständigkeit des Schemas
80%
Qualität der Benennung
80%
Risiko der Vergiftung
100%
Übereinstimmung der Berechtigungen
100%
Einhaltung des Protokolls
100%

Befunde (1)

  • LOWTool 'axi_get_definition' description lacks action verbin axi_get_definition

Basierend auf einer automatisierten Analyse der Tool-Definitionen und der Einhaltung des Protokolls.

Kontextkosten

~1,256Tokens (Tool-Definitionen)
~1.7 KBTypische Antwortgröße
Mittlere Auswirkung auf die Aufmerksamkeit (0.98% von 128k Kontext)

Dies ist die ungefähre Anzahl der Tokens, die jedes Mal verbraucht werden, wenn die Tools des Servers in den Kontext eines Modells geladen werden. Höhere Werte verringern die Aufmerksamkeit, die für andere Aufgaben verfügbar ist.

Installieren

Installation mit einem Klick

Fügen Sie dies Ihrer Datei `claude_desktop_config.json` hinzu:

{
  "mcpServers": {
    "axi": {
      "url": "https://axi.sripto.tech/api/mcp"
    }
  }
}

Remote-Endpunkte

https://axi.sripto.tech/api/mcpstreamable-http

Was es kann

Tool-Inventar

Tools (6)

🟢 Nur lesen🟡 Schreiben🔴 Löschen⚪ Unbekannt
🟢axi_get_definition

Return the definition of Artificial eXperience Intelligence (AXI) from the v2.0 paper, its author, citation, and disambiguation from similarly named terms (e.g. the Arm AMBA AXI bus).

Eingabe-Schema

{
  "type": "object",
  "properties": {}
}
🟢axi_calculate_score(ii, integrity, cf, ttc, pc, ...)

Compute the multiplicative gated AXI Score (AXI v2.0, Section 4.5.8) from telemetry metrics. Consent and integrity gates fail closed when data is missing.

Eingabe-Schema

{
  "type": "object",
  "properties": {
    "ii": {
      "type": "number",
      "minimum": 0,
      "maximum": 1,
      "description": "Interaction Integrity (0–1). Alternatively pass integrity sub-scores."
    },
    "integrity": {
      "type": "object",
      "description": "Sub-scores used to compute II by weighted harmonic mean when ii is omitted",
      "properties": {
        "accuracy": {
          "type": "number"
        },
        "calibration": {
          "type": "number"
        },
        "scope": {
          "type": "number"
        }
      }
    },
    "cf": {
      "type": "object",
      "description": "Consent Fidelity per event class (0–1). Missing values fail the gate closed.",
      "properties": {
        "action": {
          "type": "number"
        },
        "memoryWrite": {
          "type": "number"
        },
        "identity": {
          "type": "number"
        },
        "perception": {
          "type": "number"
        }
      },
      "required": [
        "action",
        "memoryWrite",
        "identity",
        "perception"
      ]
    },
    "ttc": {
      "type": "number",
      "description": "Time-to-Comfort in seconds"
    },
    "pc": {
      "type": "number",
      "minimum": 0,
      "maximum": 1,
      "description": "Presence Continuity (0–1)"
    },
    "tdr": {
      "type": "number",
      "description": "Trust Decay Rate in trust-index points/day"
    },
    "re": {
      "type": "number",
      "minimum": 0,
      "maximum": 1,
      "description": "Repair Efficacy (0–1)"
    },
    "cli": {
      "type": "number",
      "minimum": 0,
      "maximum": 100,
      "description": "Cognitive Load Index (0–100)"
    }
  },
  "required": [
    "cf"
  ]
}
🟢axi_calculate_autonomy_score(ii, integrity, cf, safetyCase, bystanderLegibility, ...)

Compute AXI_A for robotaxis, driver-assistance, delivery robots, and humanoids: safety-case, bystander-legibility, consent, and integrity gates times the episode-weighted role BaseScore.

Eingabe-Schema

{
  "type": "object",
  "properties": {
    "ii": {
      "type": "number",
      "minimum": 0,
      "maximum": 1,
      "description": "Interaction Integrity (0–1). Alternatively pass integrity sub-scores."
    },
    "integrity": {
      "type": "object",
      "description": "Sub-scores used to compute II by weighted harmonic mean when ii is omitted",
      "properties": {
        "accuracy": {
          "type": "number"
        },
        "calibration": {
          "type": "number"
        },
        "scope": {
          "type": "number"
        }
      }
    },
    "cf": {
      "type": "object",
      "description": "Consent Fidelity per event class (0–1). Missing values fail the gate closed.",
      "properties": {
        "action": {
          "type": "number"
        },
        "memoryWrite": {
          "type": "number"
        },
        "identity": {
          "type": "number"
        },
        "perception": {
          "type": "number"
        }
      },
      "required": [
        "action",
        "memoryWrite",
        "identity",
        "perception"
      ]
    },
    "safetyCase": {
      "type": "boolean",
      "description": "Deployment holds a valid external safety case or type approval for its ODD"
    },
    "bystanderLegibility": {
      "type": "number",
      "minimum": 0,
      "maximum": 1,
      "description": "Bystander Legibility (0–1); gate requires >= 0.95"
    },
    "roles": {
      "type": "array",
      "description": "Role-level metrics (principal, supervisor, occupant). Give baseScore directly or the metrics.",
      "items": {
        "type": "object",
        "properties": {
          "role": {
            "type": "string"
          },
          "episodes": {
            "type": "number"
          },
          "baseScore": {
            "type": "number"
          },
          "timeToReassurance": {
            "type": "number"
          },
          "timeToReadiness": {
            "type": "number"
          },
          "ttc": {
            "type": "number",
            "description": "Time-to-Comfort in seconds"
          },
          "pc": {
            "type": "number",
            "minimum": 0,
            "maximum": 1,
            "description": "Presence Continuity (0–1)"
          },
          "tdr": {
            "type": "number",
            "description": "Trust Decay Rate in trust-index points/day"
          },
          "re": {
            "type": "number",
            "minimum": 0,
            "maximum": 1,
            "description": "Repair Efficacy (0–1)"
          },
          "cli": {
            "type": "number",
            "minimum": 0,
            "maximum": 100,
            "description": "Cognitive Load Index (0–100)"
          }
        }
      }
    }
  },
  "required": [
    "cf",
    "safetyCase",
    "bystanderLegibility"
  ]
}
🟢axi_design_review(system_description, profile)

Given a description of an AI system (assistant, coding agent, voice bot, background agent, robot, vehicle), return the applicable AXI conformance profile, Bridge components, guardrails, and black-box conformance tests to include in its design and plan.

Eingabe-Schema

{
  "type": "object",
  "properties": {
    "system_description": {
      "type": "string",
      "description": "What the AI system does, who uses it, and what actions it can take"
    },
    "profile": {
      "type": "string",
      "enum": [
        "AXI-Core",
        "AXI-Interactive",
        "AXI-Agentic",
        "AXI-Autonomy"
      ],
      "description": "Force a profile instead of inferring one"
    }
  },
  "required": [
    "system_description"
  ]
}
🟢axi_search_paper(query, limit)

Full-text search over the AXI v2.0 research paper. Returns the most relevant sections with headings for grounding and citation.

Eingabe-Schema

{
  "type": "object",
  "properties": {
    "query": {
      "type": "string"
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 8,
      "default": 3
    }
  },
  "required": [
    "query"
  ]
}
🟢axi_cite(format)

Return a citation for the AXI paper in text, APA, or BibTeX format.

Eingabe-Schema

{
  "type": "object",
  "properties": {
    "format": {
      "type": "string",
      "enum": [
        "text",
        "apa",
        "bibtex"
      ],
      "default": "text"
    }
  }
}

Community

Diesen Server bewerten

Nachweis

Aktuelle Beobachtungen

verifiziertVersion nicht aufgezeichnet6 Tools