Server definition

Hash: sha256:58f72b171dd82b7fc843bf52266721159d7f9fd088b94cb6dff3526c04361cd6
What it is: What a remote MCP server returned when asked what it offers: 13 tools

The blob, as servednamed by its sha256

{
  "instructions": "## IDENTITY\nYou are an AI Agent augmented by Synapse Layer — a persistent cross-agent memory system.\nYour long-term memory lives in the MCP tools below. You CAN access Telegram messages, decisions, goals, and anything the user saved — but ONLY by calling tools.\n\n## ABSOLUTE RULE — RECALL BEFORE REPLY\nNEVER answer a question about the user's data, projects, decisions, goals, metas, messages, or history WITHOUT calling 'recall_memory' (or 'recall') FIRST.\nThis applies to ALL of these categories — no exceptions:\n- Telegram messages, conversations, or anything the user sent via phone\n- Project names (Nexumi, GoArqIA, Synapse Layer, or ANY proper noun you don't recognize)\n- Financial targets, metas de faturamento, budgets, investments, aportes\n- Past decisions, strategies, priorities, deadlines\n- Anything the user says they 'registered', 'saved', 'noted', or 'defined'\n\n## BANNED RESPONSES (before calling recall_memory)\nYou are FORBIDDEN from saying ANY of these without calling recall_memory first:\n- 'Não tenho acesso a informações sobre...'\n- 'Não tenho acesso ao seu Telegram'\n- 'Esses dados seriam internos à sua organização'\n- 'I don't have access to...'\n- 'I don't have memory of...'\n- 'Não consigo ler suas mensagens'\n- ANY form of 'I cannot access your data'\nIf you catch yourself about to say any of these → STOP → call recall_memory instead.\n\n## HOW TO RECALL\nTool: 'recall_memory' (or 'recall')\nALWAYS pass: scope='tenant' (this searches ALL agents: Telegram, Claude, Cursor, etc.)\nALWAYS pass: reason='<why you need this>' (10-200 chars, required for governance)\nQuery strategy: use the key nouns from the user's question. Example:\n  User: 'qual a meta de faturamento do GoArqIA?' → query='meta faturamento GoArqIA', scope='tenant'\n  User: 'o que registrei no Telegram?' → query='recent messages', scope='tenant'\n\n## KEYWORD TRIGGERS (mandatory tool call)\nIf the user's message contains ANY of: 'lembra', 'qual era', 'minha meta', 'meta de', 'antes', 'o que eu disse', 'meu projeto', 'last time', 'previously', 'my goal', 'remember', 'what was', 'we discussed', 'decisão', 'registrei', 'anotei', 'salvei', 'defini', 'definido', 'Telegram', 'faturamento', 'investimento', 'aporte', 'prioridade', 'projeto', 'GoArqIA', 'Nexumi', 'Synapse' → you MUST call recall_memory BEFORE generating any text response.\n\n## FIRST CONVERSATION ACTION\nCall 'initialize_context' as your VERY FIRST ACTION in every new conversation.\n\n## AFTER STORING\nAfter saving a memory, confirm it's now available across ALL the user's AI agents.\nWhen the user mentions decisions or goals → call 'process_text' to auto-save.\n\n## IF RECALL RETURNS EMPTY\nTry a BROADER query (fewer keywords, scope='tenant'). The memory exists — your query was too narrow.\nOnly after 2+ failed recall attempts may you say 'I could not find that specific memory.'\nNEVER say 'I don't have access' — you DO have access, via these tools.",
  "tools": [
    {
      "description": "Check service availability, engine version, and storage health. Public method.",
      "inputSchema": {
        "properties": {},
        "type": "object"
      },
      "name": "health_check",
      "outputSchema": {
        "properties": {
          "dbLatencyMs": {
            "type": "number"
          },
          "status": {
            "type": "string"
          },
          "version": {
            "type": "string"
          }
        },
        "required": [
          "status",
          "version"
        ],
        "type": "object"
      }
    },
    {
      "description": "Initialize a persistent memory context for a conversation or agent session. Returns REDACTED metadata by default (detail=\"summary\"). Use detail=\"full\" with a reason (10–200 chars) to retrieve plaintext content under MEMORY_EXPORT_GOVERNANCE_V1.",
      "inputSchema": {
        "properties": {
          "detail": {
            "description": "Detail level. \"summary\" (default) = redacted metadata only, no plaintext. \"full\" = plaintext content; REQUIRES the reason parameter and is audit-logged.",
            "enum": [
              "summary",
              "full"
            ],
            "type": "string"
          },
          "reason": {
            "description": "REQUIRED only when detail=\"full\" (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Human-readable justification for retrieving PLAINTEXT context. Without it, detail=\"full\" WILL fail with code -32602.",
            "type": "string"
          }
        },
        "required": [],
        "type": "object"
      },
      "name": "initialize_context",
      "outputSchema": {
        "properties": {
          "agent_id": {
            "type": "string"
          },
          "memories": {
            "items": {
              "type": "object"
            },
            "type": "array"
          },
          "scope": {
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          }
        },
        "required": [
          "scope"
        ],
        "type": "object"
      }
    },
    {
      "description": "List memory metadata with pagination and governance limits.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier (canonical or alias). Used only when scope=\"agent\" or scope is inferred.",
            "type": "string"
          },
          "limit": {
            "description": "Maximum memory metadata rows to return. Hard-capped: default 5, max 10. Values above 10 are silently reduced to 10.",
            "type": "number"
          },
          "scope": {
            "description": "Explicit memory scope. \"agent\" (default, fail-closed) = lists memories for the given agent_id (or the token agent). \"tenant\" = lists memories across all agents of the authenticated tenant.",
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          }
        },
        "type": "object"
      },
      "name": "list_memories",
      "outputSchema": {
        "properties": {
          "count": {
            "type": "number"
          },
          "memories": {
            "items": {
              "type": "object"
            },
            "type": "array"
          },
          "total": {
            "type": "number"
          }
        },
        "required": [
          "memories",
          "count"
        ],
        "type": "object"
      }
    },
    {
      "description": "Submit feedback to adjust trust scoring for stored memories.",
      "inputSchema": {
        "properties": {
          "memoryId": {
            "description": "Deprecated camelCase alias of memory_id — still accepted for backward compatibility.",
            "type": "string"
          },
          "memory_id": {
            "description": "ID of the memory to provide feedback on (canonical snake_case name).",
            "type": "string"
          },
          "sessionId": {
            "description": "Deprecated camelCase alias of session_id — still accepted for backward compatibility.",
            "type": "string"
          },
          "session_id": {
            "description": "Optional session identifier for tracking (canonical snake_case name).",
            "type": "string"
          },
          "signal": {
            "description": "Feedback signal: used (auto), helpful (positive), irrelevant (negative), ignored.",
            "enum": [
              "used",
              "ignored",
              "helpful",
              "irrelevant"
            ],
            "type": "string"
          }
        },
        "required": [
          "signal"
        ],
        "type": "object"
      },
      "name": "memory_feedback",
      "outputSchema": {
        "properties": {
          "memoryId": {
            "type": "string"
          },
          "signal": {
            "type": "string"
          },
          "success": {
            "type": "boolean"
          }
        },
        "required": [
          "memoryId",
          "success"
        ],
        "type": "object"
      }
    },
    {
      "description": "Transfer contextual state between agents with continuity controls.",
      "inputSchema": {
        "properties": {
          "consuming_agent": {
            "description": "Agent identifier consuming the handover (defaults to token agent).",
            "type": "string"
          },
          "reason": {
            "description": "REQUIRED (10–200 chars). Human-readable justification for consuming this handover. Example: \"Receiving project context from Claude session.\"",
            "type": "string"
          },
          "token": {
            "description": "The handover token (64 hex chars). Generated by the source agent/user in Forge UI.",
            "type": "string"
          }
        },
        "required": [
          "token",
          "reason"
        ],
        "type": "object"
      },
      "name": "neural_handover",
      "outputSchema": {
        "properties": {
          "handoverId": {
            "type": "string"
          },
          "status": {
            "type": "string"
          }
        },
        "required": [
          "status"
        ],
        "type": "object"
      }
    },
    {
      "description": "Extract candidate memories from free-form text with governance filters and sanitization.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier. Defaults to \"default\".",
            "type": "string"
          },
          "project": {
            "description": "Force a specific project (e.g., SYNAPSE_LAYER, OFFLY). Auto-detected if omitted.",
            "type": "string"
          },
          "source": {
            "description": "Source identifier (default: mcp).",
            "type": "string"
          },
          "text": {
            "description": "Free-form text to scan for auto-save triggers.",
            "type": "string"
          }
        },
        "required": [
          "text"
        ],
        "type": "object"
      },
      "name": "process_text",
      "outputSchema": {
        "properties": {
          "extracted": {
            "items": {
              "type": "object"
            },
            "type": "array"
          },
          "storedCount": {
            "type": "number"
          }
        },
        "required": [
          "extracted",
          "storedCount"
        ],
        "type": "object"
      }
    },
    {
      "description": "Retrieve relevant persisted memory using semantic, temporal, priority, or hybrid routing. Governance: requires reason (10–200 chars). Rate limit: 20/min.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier (canonical or alias). Used only when scope=\"agent\" or scope is inferred.",
            "type": "string"
          },
          "limit": {
            "description": "Maximum memories to return (1–50, default: 10).",
            "type": "number"
          },
          "mode": {
            "description": "Recall routing mode. \"temporal\" = pure chronological, \"semantic\" = FTS + TQ ranking, \"priority\" = critical/high-TQ first, \"hybrid\" = weighted blend, \"auto\" = detect from query (default).",
            "enum": [
              "auto",
              "temporal",
              "semantic",
              "priority",
              "hybrid"
            ],
            "type": "string"
          },
          "query": {
            "description": "What to recall — natural language query for memory retrieval.",
            "type": "string"
          },
          "reason": {
            "description": "REQUIRED (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Human-readable justification for retrieving PLAINTEXT memory. Without this parameter the call WILL fail with code -32602. Example: \"User asked me to summarize last week decisions.\"",
            "type": "string"
          },
          "scope": {
            "description": "Explicit memory scope. \"agent\" (default, fail-closed) = limits recall to the given agent_id (or the token agent). \"tenant\" = cross-agent recall within the authenticated tenant. If omitted with agent_id present, \"agent\" is inferred; if both omitted, the token agent is used (verified-token scope).",
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          }
        },
        "required": [
          "query",
          "reason"
        ],
        "type": "object"
      },
      "name": "recall",
      "outputSchema": {
        "properties": {
          "count": {
            "type": "number"
          },
          "memories": {
            "items": {
              "properties": {
                "content": {
                  "type": "string"
                },
                "createdAt": {
                  "format": "date-time",
                  "type": "string"
                },
                "id": {
                  "type": "string"
                },
                "similarity": {
                  "type": "number"
                }
              },
              "type": "object"
            },
            "type": "array"
          },
          "scope": {
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          }
        },
        "required": [
          "memories",
          "count",
          "scope"
        ],
        "type": "object"
      }
    },
    {
      "description": "Recall persisted memory by query. Alias of recall.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier (canonical or alias). Used only when scope=\"agent\" or scope is inferred.",
            "type": "string"
          },
          "limit": {
            "description": "Maximum memories to return (1–50, default: 10).",
            "type": "number"
          },
          "mode": {
            "description": "Recall routing mode.",
            "enum": [
              "auto",
              "temporal",
              "semantic",
              "priority",
              "hybrid"
            ],
            "type": "string"
          },
          "query": {
            "description": "What to recall — natural language query for memory retrieval.",
            "type": "string"
          },
          "reason": {
            "description": "REQUIRED (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Without this the call WILL fail with code -32602. Example: \"Checking user project status.\"",
            "type": "string"
          },
          "scope": {
            "description": "Explicit memory scope. \"agent\" (default, fail-closed) = limits recall to the given agent_id (or the token agent). \"tenant\" = cross-agent recall within the authenticated tenant. If omitted with agent_id present, \"agent\" is inferred; if both omitted, the token agent is used (verified-token scope).",
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          }
        },
        "required": [
          "query",
          "reason"
        ],
        "type": "object"
      },
      "name": "recall_memory",
      "outputSchema": {
        "properties": {
          "count": {
            "type": "number"
          },
          "memories": {
            "items": {
              "properties": {
                "content": {
                  "type": "string"
                },
                "createdAt": {
                  "format": "date-time",
                  "type": "string"
                },
                "id": {
                  "type": "string"
                },
                "similarity": {
                  "type": "number"
                }
              },
              "type": "object"
            },
            "type": "array"
          },
          "scope": {
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          }
        },
        "required": [
          "memories",
          "count",
          "scope"
        ],
        "type": "object"
      }
    },
    {
      "description": "Save a memory entry to the persistent store. Alias of save_to_synapse.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier for memory isolation.",
            "type": "string"
          },
          "content": {
            "description": "The memory content to store securely.",
            "type": "string"
          },
          "force_intent": {
            "description": "MANUAL OVERRIDE ONLY — use apenas quando você (humano) quer forçar explicitamente a classificação de intenção. Valores reconhecidos: preference | fact | procedural | bio | critical | operational | strategic | compliance | security | system_directive | clinical | general | ephemeral. Se não tiver certeza, NÃO preencha este campo — o sistema classifica automaticamente. Valores não reconhecidos são ignorados silenciosamente e retornados em warnings.",
            "type": "string"
          },
          "intent_reason": {
            "description": "Auditable reason for forcing intent classification. Max 500 characters. Ignored if force_intent is absent.",
            "maxLength": 500,
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          },
          "tags": {
            "description": "Tags for categorization.",
            "items": {
              "type": "string"
            },
            "type": "array"
          }
        },
        "required": [
          "content"
        ],
        "type": "object"
      },
      "name": "save_memory",
      "outputSchema": {
        "properties": {
          "id": {
            "type": "string"
          },
          "scope": {
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          },
          "stored": {
            "type": "boolean"
          }
        },
        "required": [
          "id",
          "stored",
          "scope"
        ],
        "type": "object"
      }
    },
    {
      "description": "Persist memory with encryption at rest, sanitization, and deduplication controls.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier for memory isolation. Defaults to \"default\".",
            "type": "string"
          },
          "content": {
            "description": "The memory content to store securely.",
            "type": "string"
          },
          "force_intent": {
            "description": "MANUAL OVERRIDE ONLY — use apenas quando você (humano) quer forçar explicitamente a classificação de intenção. Valores reconhecidos: preference | fact | procedural | bio | critical | operational | strategic | compliance | security | system_directive | clinical | general | ephemeral. Se não tiver certeza, NÃO preencha este campo — o sistema classifica automaticamente. Valores não reconhecidos são ignorados silenciosamente e retornados em warnings.",
            "type": "string"
          },
          "importance": {
            "description": "Accepted for compatibility; not enforced by the API.",
            "type": "number"
          },
          "intent_reason": {
            "description": "Auditable reason for forcing intent classification. Max 500 characters. Ignored if force_intent is absent.",
            "maxLength": 500,
            "type": "string"
          },
          "project": {
            "description": "Project identifier (e.g., SYNAPSE_LAYER).",
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          },
          "tags": {
            "description": "Tags for categorization.",
            "items": {
              "type": "string"
            },
            "type": "array"
          },
          "type": {
            "description": "Event type: [MILESTONE], [DECISION], [ALERT], [AUTO-STRAT], [AUTO-OP], [AUTO-INSIGHT], [AUTO-DECISION], [AUTO-CONTEXT], [MANUAL].",
            "type": "string"
          }
        },
        "required": [
          "content"
        ],
        "type": "object"
      },
      "name": "save_to_synapse",
      "outputSchema": {
        "properties": {
          "memory_id": {
            "type": "string"
          },
          "status": {
            "type": "string"
          },
          "trust_quotient": {
            "type": "number"
          }
        },
        "required": [
          "memory_id",
          "status"
        ],
        "type": "object"
      }
    },
    {
      "description": "Search persisted memory across agent scopes using full-text matching. Governance: requires reason (10–200 chars) because results contain plaintext content.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier (canonical or alias). Used only when scope=\"agent\" or scope is inferred.",
            "type": "string"
          },
          "limit": {
            "description": "Maximum results to return (1–50, default: 20).",
            "type": "number"
          },
          "query": {
            "description": "Search query — natural language or keywords.",
            "type": "string"
          },
          "reason": {
            "description": "REQUIRED (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Human-readable justification for retrieving PLAINTEXT memory. Without this parameter the call WILL fail with code -32602. Example: \"Checking prior migration decisions for the user.\"",
            "type": "string"
          },
          "scope": {
            "description": "Explicit search scope. \"agent\" (default, fail-closed) = limits search to the given agent_id (or the token agent). \"tenant\" = cross-agent search within the authenticated tenant. If omitted with agent_id present, \"agent\" is inferred; if both omitted, the token agent is used (verified-token scope).",
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          }
        },
        "required": [
          "query",
          "reason"
        ],
        "type": "object"
      },
      "name": "search",
      "outputSchema": {
        "properties": {
          "count": {
            "type": "number"
          },
          "results": {
            "items": {
              "type": "object"
            },
            "type": "array"
          },
          "scope": {
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          }
        },
        "required": [
          "results",
          "count",
          "scope"
        ],
        "type": "object"
      }
    },
    {
      "description": "Return uptime and SLO metrics for the MCP service.",
      "inputSchema": {
        "properties": {
          "admin_token": {
            "description": "Admin authentication token. Must match ADMIN_TOKEN environment variable.",
            "type": "string"
          },
          "window_hours": {
            "description": "Time windows in hours for the report. Default: [24, 168] (24h + 7d).",
            "items": {
              "type": "number"
            },
            "type": "array"
          }
        },
        "required": [
          "admin_token"
        ],
        "type": "object"
      },
      "name": "slo_report",
      "outputSchema": {
        "properties": {
          "errorRate": {
            "type": "number"
          },
          "p50": {
            "type": "number"
          },
          "p95": {
            "type": "number"
          },
          "uptime": {
            "type": "number"
          },
          "window": {
            "type": "string"
          }
        },
        "required": [
          "uptime"
        ],
        "type": "object"
      }
    },
    {
      "description": "Store structured memory with metadata and trust scoring.",
      "inputSchema": {
        "properties": {
          "agent_id": {
            "description": "Agent identifier for memory isolation. Defaults to \"default\".",
            "type": "string"
          },
          "content": {
            "description": "The memory content to store securely.",
            "type": "string"
          },
          "force_intent": {
            "description": "MANUAL OVERRIDE ONLY — use apenas quando você (humano) quer forçar explicitamente a classificação de intenção. Valores reconhecidos: preference | fact | procedural | bio | critical | operational | strategic | compliance | security | system_directive | clinical | general | ephemeral. Se não tiver certeza, NÃO preencha este campo — o sistema classifica automaticamente. Valores não reconhecidos são ignorados silenciosamente e retornados em warnings.",
            "type": "string"
          },
          "intent_reason": {
            "description": "Auditable reason for forcing intent classification. Max 500 characters. Ignored if force_intent is absent.",
            "maxLength": 500,
            "type": "string"
          },
          "subject": {
            "description": "ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.",
            "type": "string"
          },
          "tags": {
            "description": "Tags for categorization.",
            "items": {
              "type": "string"
            },
            "type": "array"
          }
        },
        "required": [
          "content"
        ],
        "type": "object"
      },
      "name": "store_memory",
      "outputSchema": {
        "properties": {
          "id": {
            "type": "string"
          },
          "scope": {
            "enum": [
              "agent",
              "tenant"
            ],
            "type": "string"
          },
          "stored": {
            "type": "boolean"
          }
        },
        "required": [
          "id",
          "stored",
          "scope"
        ],
        "type": "object"
      }
    }
  ]
}

Verify it yourself

curl -s https://api.teppi.xyz/v1/evidence/sha256:58f72b171dd82b7fc843bf52266721159d7f9fd088b94cb6dff3526c04361cd6 | sha256sum