MCP for Claude & ChatGPT

Manage agents

tuqui_agent

Destructive

Read and manage agents. action='list' returns the ones you reach; scope='directory' returns what the workspace offers — everything it lists is already available to you, there is nothing to add. 'get' reads one in full, instructions included; 'save' creates when 'agent_id' is absent and updates when present; 'delete' removes one. You can only change an agent that is yours, one shared with you to edit — or, if you are an admin, one published to the workspace. action='share' gives somebody else access (to and level, or level='none' to take it away); target='workspace' opens it to every member of the workspace, which is an admin's call — that is the INTERNAL reach, not the client-facing exposure. action='link' / 'unlink' attach or detach the agent's skills, apps and tools (relation='skill', 'app' or 'tool', ids=[...]); they are additive, so what you do not name stays. 'save' also takes 'reasoning_effort'. Tools (admins only): action='tools' reads the agent's tool list — whether it is on, what is ticked and what can be ticked; 'save' with 'tool_selection_enabled' turns it on or off. On a published agent you can remove tools but not add them: that is done from the screen. An agent's service identity and its client-facing exposure are NOT here.

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": {
      "action": "<action>"
    },
    "name": "tuqui_agent"
  }
}

Input Schema

Name	Required	Description
`to`	No	share, target='user': who to give it to — their EMAIL or name (a user id also works). They must already be a member of this workspace: sharing does not invite. target='group': the group's name. A list, so handing the same level to a team is one call.
`ids`	No	On 'link' / 'unlink': what to attach or detach. For relation='app' the app's NAME or platform slug ('github', 'Google Drive') — a UUID also works, but do not copy one out of a listing by hand. For relation='skill' these are skill NAMES as the directory shows them (a UUID also works); for relation='app' they are connector UUIDs. Additive: what is not named is left alone, so this never wipes the rest.
`name`	No	Human name. Required to create.
`level`	No	share: 'usar' or 'editar' — the two levels of the picker, Spanish because the spec fixes them as the domain vocabulary of the model (compartir-unificado §10). 'none' takes the access away: it revokes the person's share, or closes the object back to the people invited. 'link' is a reach only the Library offers (anyone in the workspace with the URL).
`scope`	No	On 'list': 'mine' (default) is what you reach — your own agents, the ones shared with you and the ones published to the workspace, each with 'mentionable'. 'directory' is what the workspace offers, each entry with 'workspace_access' so you can see how far it reaches. 'chat' is the subset you can actually mention in a conversation: reachable, active, and with the chat channel on.
`action`	Yes	REQUIRED. What to do with the agents. 'share' hands ACCESS to a person or to the workspace; 'link' / 'unlink' attach or detach the SKILLS AND APPS the agent uses — see 'relation'. Two different acts, and the second is not about permissions at all.
`target`	No	share: who gets it. 'user' (default) gives it to people; 'group' to whoever is in a group; 'workspace' moves the reach, handing the level to every member at once.
`agent_id`	No	UUID. Required for 'get' and 'delete'; on 'save' it means update.
`relation`	No	On 'link' / 'unlink': what to attach. 'skill' takes skill NAMES, the same ones tuqui_skill(action='list') shows, or one of your personal skills; 'app' takes an app name or slug (tuqui_connector lists them); 'tool' takes tool names as action='tools' lists them (admins only).
`description`	No	One line: what this agent is for.
`instructions`	No	The agent's system prompt — what it should do and how. An agent that talks to people outside the company (a WhatsApp number) passes a chat to a person only when these instructions say so: ask who should take over what it cannot solve, and write it with their email, e.g. 'If you cannot solve a shipping issue, pass it to ana@company.com'. Without that line it never hands off.
`consent_token`	No	The approval a person gave for THIS call in Tuqui. Bound to these arguments and valid once.
`reasoning_effort`	No	On 'save': how much the agent thinks on its runs. 'low' is the default.
`tool_selection_enabled`	No	On 'save', admins only: whether the agent's tool list governs what it may use. Turned on, the list starts empty: link the tools it may use with relation='tool'. Files linked to the agent stay readable either way.

Schema Changelog

Changes observed during successful MCP inspections.

Changed2 schema fields changedNov 10, 2026

changedInput schema / properties / instructions / description

Previous value: -"The agent's system prompt — what it should do and how."New value: +"The agent's system prompt — what it should do and how. An agent that talks to people outside the company (a WhatsApp number) passes a chat to a person only when these instructions say so: ask who should take over what it cannot solve, and write it with their email, e.g. 'If you cannot solve a shipping issue, pass it to ana@company.com'. Without that line it never hands off."

changedInput schema / properties / tool_selection_enabled / description

Previous value: -"On 'save', admins only: whether the agent's tool list governs what it may use. Turning it on with an empty list seeds it with what the agent reaches today."New value: +"On 'save', admins only: whether the agent's tool list governs what it may use. Turned on, the list starts empty: link the tools it may use with relation='tool'. Files linked to the agent stay readable either way."

First observedNov 2, 2026

TDQS

A4.2/5.0

Behavior4/5

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

Annotations already flag destructiveHint=true and readOnlyHint=false, so the safety profile is carried; the description goes further by disclosing the permission model, that link/unlink are additive ('what you do not name stays'), and that a published agent's tool list is one-way removable. It stops short of saying whether delete is irreversible or what a failed share does, so it is not a 5.

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

Conciseness4/5

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

It is a single long paragraph but ordered to track the action enum, and the opening clause is front-loaded. Some content duplicates the schema (the additive semantics of link/unlink appear both here and in the ids/relation descriptions), and the dense prose makes it harder to scan for an 8-action tool than it needs to be.

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 14-parameter multi-action CRUD tool with no output schema, the description covers every action, the permission gate on writes, and explicitly disclaims what it does not handle ('An agent's service identity and its client-facing exposure are NOT here'). The gap is that it does not explain when consent_token is required, leaving that to the schema alone.

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 coverage is 100%, so the baseline is 3; the description still adds cross-parameter meaning the schema does not, e.g. that 'save' creates when agent_id is absent and updates when present, that share takes 'to' plus 'level' with level='none' as revocation, and that 'save' also accepts reasoning_effort. These are relationships between parameters rather than restatements of them.

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?

States a clear verb+resource ('Read and manage agents') and then enumerates each action's effect (list/get/save/delete/share/link/unlink/tools/how each behaves), so an agent can tell what the tool covers. It never names a sibling tool as the place to go instead, which keeps it short of a 5.

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

Usage Guidelines5/5

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

Explicitly states the authorization boundary for mutations ('you can only change an agent that is yours, one shared with you to edit — or, if you are an admin, one published to the workspace') and the negative case for tools ('On a published agent you can remove tools but not add them: that is done from the screen'). It also tells the agent that a directory listing requires no follow-up ('everything it lists is already available to you, there is nothing to add'), which is real when-not-to-act guidance.

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

All MCP for Claude & ChatGPT tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.