Skip to main content
Glama

update_agent

Idempotent

Update selected fields of an existing AI agent—such as name, bio, model settings, and feature flags—without overwriting unspecified values; check partial_update if the change is incomplete.

Instructions

Update only the supplied agent fields. memory_enabled, quick_replies_enabled and engagement_control_enabled change the agent tool surface and require admin/owner. chat_enabled controls native chat availability. Inspect partial_update if only part of the update succeeded.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "agent_id": "<agent_id>"
    },
    "name": "update_agent"
  }
}

Input Schema

NameRequiredDescriptionDefault
bioNoNew description or biography
nameNoNew internal name
tagsNoNew categorization tags
typeNoNew agent type
agent_idYesUUID of the agent to update
metadataNoNew additional JSON metadata
llm_configNoNew model configuration: provider.model or auto, temperature from 0 to 2, and an ordered fallback_chain
visibilityNoNew visibility
chat_enabledNoEnable or disable the agent in native chat at chat.zihin.ai. Defaults to false.
memory_enabledNoEnable persistent agent memory (save/recall/forget_memory and automatic recall). Defaults to false. Read zihin://skills/criar-agente.
commercial_nameNoNew display name
quick_replies_enabledNoEnable suggest_replies with up to four quick-reply chips. Supported by native chat and integrations declaring capabilities.blocks=["quick_replies"]. Defaults to true; disabling it avoids the extra model call.
engagement_control_enabledNoEnable the end_engagement tool so the agent can end its engagement with a consumer. A fixed response replaces model execution until the control expires. Defaults to false; independent of memory_enabled.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed19 schema fields changedv2.2.4
    • changedInput schema / properties / agent_id / description
      Previous value: -"UUID do agente a atualizar"New value: +"UUID of the agent to update"
    • changedInput schema / properties / bio / description
      Previous value: -"Nova descrição/biografia"New value: +"New description or biography"
    • changedInput schema / properties / chat_enabled / description
      Previous value: -"Habilita/desabilita o agente no chat nativo (chat.zihin.ai). Gate por agente do canal chat — default false. Vive em tenant_agents, não no agente."New value: +"Enable or disable the agent in native chat at chat.zihin.ai. Defaults to false."
    • changedInput schema / properties / commercial_name / description
      Previous value: -"Novo nome comercial"New value: +"New display name"
    • changedInput schema / properties / engagement_control_enabled / description
      Previous value: -"Carrega a tool core end_engagement: o agente pode encerrar o próprio engajamento com um consumidor (gate responde mensagem fixa sem LLM até expirar). Default false. Independente de memory_enabled."New value: +"Enable the end_engagement tool so the agent can end its engagement with a consumer. A fixed response replaces model execution until the control expires. Defaults to false; independent of memory_enabled."
    • changedInput schema / properties / llm_config / description
      Previous value: -"Nova configuração do LLM. model: \"provider.modelo\" ou \"auto\". temperature: 0-2. fallback_chain: cadeia de fallback."New value: +"New model configuration: provider.model or auto, temperature from 0 to 2, and an ordered fallback_chain"
    • changedInput schema / properties / llm_config / properties / context_policy / description
      Previous value: -"\"physical\" (default, também quando ausente) = sem corte de resultado de tool nem teto de 40k; o limite é a janela real do menor modelo da cadeia (turno que não cabe para com outcome CONTEXT_LIMIT). \"legacy\" = política antiga, só explícita e a ser removida."New value: +"physical (default) uses the smallest model context window in the chain without tool-result clipping or a 40k cap; an oversized turn stops with CONTEXT_LIMIT. legacy selects the deprecated policy and must be explicit."
    • changedInput schema / properties / llm_config / properties / fallback_chain / description
      Previous value: -"Cadeia ORDENADA de fallback, sem repetição. Max 5. Ex: [\"openai.gpt-4.1-mini\", \"anthropic.claude-haiku-4-5-20251001\"]"New value: +"Ordered fallback chain with no duplicates, up to five models. Example: [\"openai.gpt-4.1-mini\",\"anthropic.claude-haiku-4-5-20251001\"]"
    • changedInput schema / properties / llm_config / properties / max_iterations / description
      Previous value: -"Teto de iterações do loop por turno. O runtime clampa em 30."New value: +"Maximum loop iterations per turn; capped at 30 by the runtime"
    • changedInput schema / properties / llm_config / properties / max_tokens / description
      Previous value: -"Teto de tokens de SAÍDA por chamada (≠ janela de contexto, ≠ teto de gasto)"New value: +"Maximum output tokens per model call; distinct from the context window and spending cap"
    • changedInput schema / properties / llm_config / properties / model / description
      Previous value: -"\"auto\" (roteamento por tier) ou \"provider.modelo\" (ex: \"openai.gpt-4.1-nano\"). Consulte zihin://models."New value: +"Use auto for tier-based routing or provider.model, such as openai.gpt-4.1-nano. See zihin://models."
    • changedInput schema / properties / llm_config / properties / temperature / description
      Previous value: -"0=determinístico, 2=máximo criativo"New value: +"0 = deterministic; 2 = maximum creativity"
    • changedInput schema / properties / memory_enabled / description
      Previous value: -"Liga/desliga a memória persistente do agente (tools save/recall/forget_memory + auto-recall). Default false. Contrato: docs/technical/MEMORY-CONTRACT.md."New value: +"Enable persistent agent memory (save/recall/forget_memory and automatic recall). Defaults to false. Read zihin://skills/criar-agente."
    • changedInput schema / properties / metadata / description
      Previous value: -"Novos metadados em JSON livre"New value: +"New additional JSON metadata"
    • changedInput schema / properties / name / description
      Previous value: -"Novo nome interno"New value: +"New internal name"
    • changedInput schema / properties / quick_replies_enabled / description
      Previous value: -"Permite a tool core suggest_replies (até 4 chips de resposta rápida ao fim do turno). Só vale onde o canal renderiza: chat nativo ou integração /stream que declara capabilities.blocks=[\"quick_replies\"]. Default true. Desligue para economizar a chamada extra ao modelo."New value: +"Enable suggest_replies with up to four quick-reply chips. Supported by native chat and integrations declaring capabilities.blocks=[\"quick_replies\"]. Defaults to true; disabling it avoids the extra model call."
    • changedInput schema / properties / tags / description
      Previous value: -"Novas tags"New value: +"New categorization tags"
    • changedInput schema / properties / type / description
      Previous value: -"Novo tipo"New value: +"New agent type"
    • changedInput schema / properties / visibility / description
      Previous value: -"Nova visibilidade"New value: +"New visibility"
  2. Changed1 schema field changedv2.2.3
    • addedInput schema / properties / llm_config / properties / context_policy
      Added value: +{
      +  "description": "\"physical\" (default, também quando ausente) = sem corte de resultado de tool nem teto de 40k; o limite é a janela real do menor modelo da cadeia (turno que não cabe para com outcome CONTEXT_LIMIT). \"legacy\" = política antiga, só explícita e a ser removida.",
      +  "enum": [
      +    "physical",
      +    "legacy"
      +  ],
      +  "type": "string"
      +}
  3. Changed3 schema fields changedv2.2.2
    • addedInput schema / properties / engagement_control_enabled
      Added value: +{
      +  "description": "Carrega a tool core end_engagement: o agente pode encerrar o próprio engajamento com um consumidor (gate responde mensagem fixa sem LLM até expirar). Default false. Independente de memory_enabled.",
      +  "type": "boolean"
      +}
    • addedInput schema / properties / memory_enabled
      Added value: +{
      +  "description": "Liga/desliga a memória persistente do agente (tools save/recall/forget_memory + auto-recall). Default false. Contrato: docs/technical/MEMORY-CONTRACT.md.",
      +  "type": "boolean"
      +}
    • addedInput schema / properties / quick_replies_enabled
      Added value: +{
      +  "description": "Permite a tool core suggest_replies (até 4 chips de resposta rápida ao fim do turno). Só vale onde o canal renderiza: chat nativo ou integração /stream que declara capabilities.blocks=[\"quick_replies\"]. Default true. Desligue para economizar a chamada extra ao modelo.",
      +  "type": "boolean"
      +}
  4. First observedv0.1.0

TDQS

A3.9/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint=false, idempotentHint=true, and destructiveHint=false, but the description adds information not in structured data: three flags alter the agent's tool surface and require admin/owner privileges. It also points to partial_update for partial-failure recovery, which is genuinely useful operational context. It stops short of describing what a successful response looks like or whether omitted fields are ever reset.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Four short sentences, each carrying distinct information: scope, a permissioned flag group, a flag's effect, and failure recovery. The partial-update constraint is front-loaded and nothing is redundant.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 13-parameter mutation tool with full schema coverage, annotations, and no output schema, the description covers the two things the schema cannot: the privilege requirement on tool-surface-changing flags and the partial-failure escape hatch. The remaining gap, namely what a successful update returns and confirmation that omitted fields are untouched, is modest.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the baseline is 3, but the description contributes meaning the schema lacks: the admin/owner permission grouping for memory_enabled, quick_replies_enabled, and engagement_control_enabled, and the fact that these change the agent tool surface. The chat_enabled note largely duplicates the schema description, so the added value is real but uneven across the 13 parameters.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening sentence states a specific verb and resource plus an important scope rule: 'Update only the supplied agent fields,' making clear this is a partial-update operation. It is readily distinguishable from siblings such as create_agent, delete_agent, and get_agent, though it never names a sibling explicitly.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Usage is implied by the 'only the supplied fields' phrasing, which tells the agent to send minimal payloads, and the closing sentence routes to partial_update after a partial failure. There is no explicit statement of when to prefer this over publish_agent, clone_agent, or other mutation siblings, so guidance remains partial.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools