Koskamo

AgentStack MCP

AgentStack provides a backend ecosystem for AI agents, exposing over 700 actions via a single MCP tool named agentstack.execute. It supports hosting, storage, authentication, payments, and project management, allowing agents to autonomously manage products and data through a Work Graph loop.

Endpoint and access

No sign-in to connect and list tools. Individual tools may still need a vendor account, credentials, or permissions; check the documented prerequisites.

AgentStack MCP logo, by AgentStackCommunity listingStreamable HTTPNo sign-inEndpoint checked 2026-11-07Published by AgentStack
MCP endpoint (Streamable HTTP)
https://agentstack.tech/mcp
Vendor documentation

What its tools do

1 tool, grouped by the hints the server advertises. Hints are self-declared, not verified behavior.

Can change data (1)

1 of 1 have a description · 1 of 1 declare behavior hints · 1 of 1 with parameters document them all

Checks and signals

Each line is separate evidence, not a trust score.

  • Passed Endpoint answered the live check ·
  • Availability 1 of 1 checks passed in 30 days

What you can do with it

  • Host static sites and SPAs on AgentStack
  • Manage project authentication and storage
  • Process payments and commerce transactions
  • Execute backend operations via agentstack.execute

Source and maintenance

Community listingRegistry 1.0.0
Maintenance checked

Protocol and capabilities

Server
agentstack 0.4.26
MCP protocol negotiated in check
2025-11-25
Tools
1 advertised
Resources
4 advertised
Resource templates
2 advertised
Prompts
72 advertised
Server-advertised capabilities
{
  "tasks": {
    "cancel": {},
    "requests": {
      "tools": {
        "call": {}
      }
    }
  },
  "tools": {},
  "logging": {},
  "prompts": {},
  "resources": {},
  "completions": {}
}

Server-provided instructions

Untrusted server output shown for inspection. It is not Koskamo guidance and is excluded from knowledge search.

AgentStack MCP — one tool: agentstack.execute (POST /mcp). Bound project: agents.work_next (recipe mcp_work_loop_v1) before catalog search. discovery.search is fallback only. Authenticate first: GET /mcp/connect → /mcp/.well-known/oauth-authorize (PKCE), IDE Device Code POST /api/oauth2/device/authorize, or MCP auth.login. If this client cannot complete OAuth, use the user's API key: header X-API-Key or Authorization: Bearer <user API key> from the profile. Session probe: auth.get_profile (alias auth.me). Bind context.project_id, then agents.work_next. Packs session.v1/batch.v1: discovery.list exposes execute_budget.packs. Work Graph skill: /mcp/prompts/get?name=agentstack_work_graph (skill://agentstack/work-graph). If session_ready or mutation_allowed, do not log in again and do not switch_project. discovery.search only when packet.next_action is absent. discovery.describe for one action. preflight.check before writes. Empty lists are valid. Diagnose before repair. A timeout retries the same idempotency key. Do not call a heavy_llm action only to read state. A page at /s/{id}/ uses the hosted cookie, not operator MCP auth.login. Playbook: /mcp/prompts/get?name=agentstack_session_setup. When discovery.search returns response_mode=workflow, run that step. Do not treat catalog_page as the command. OAuth often binds ecosystem project_id=1 — override with tenant id.

Tools

1 advertised

agentstack.execute — AgentStack execute
Advertised by the server during the last successful anonymous check.
Open-world

AgentStack MCP = Work Graph control plane. After session + context.project_id: agents.work_next (mcp_work_loop_v1) -> plan_claim -> plan_execute (CAS/leases). Follow packet.next_action; catalog/discovery.search only if no packet. One-shot CRM/upload may call domain actions directly. Spine: agents.plan_get -> agents.work_next -> agents.plan_claim -> agents.plan_execute. Auth: Connect (G-A174) or Device Code, then auth.get_profile (bootstrap: include_settings false, include_subscriptions false; recipe mcp_session_setup), then context.project_id on every batch. Prompts: agentstack_session_setup, agentstack_execute_budget.

agentstack.execute input schema
{
  "type": "object",
  "required": [
    "steps"
  ],
  "properties": {
    "steps": {
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "id",
          "action"
        ],
        "properties": {
          "id": {
            "type": "string",
            "description": "Step id in batch."
          },
          "action": {
            "type": "string",
            "description": "Dotted action from /mcp/actions."
          },
          "params": {
            "type": "object",
            "description": "Action params."
          }
        },
        "description": "One catalog action."
      },
      "minItems": 1,
      "description": "Steps; one heavy LLM per sync execute."
    },
    "context": {
      "type": "object",
      "properties": {
        "project_id": {
          "type": "integer",
          "description": "Tenant project id."
        }
      },
      "description": "session.v1; project_id required."
    },
    "options": {
      "type": "object",
      "properties": {
        "goal": {
          "type": "string",
          "description": "Binds the batch to a playbook. A direct call uses execution_intent direct and omits goal. Per-step query must match this text."
        },
        "async": {
          "type": "boolean",
          "description": "Queue; poll discovery.job_status."
        },
        "envelope": {
          "enum": [
            "compact",
            "full"
          ],
          "type": "string",
          "description": "compact (default) | full — ADR MCP_EXECUTE_ENVELOPE_COMPACT_V2"
        },
        "recipe_id": {
          "type": "string",
          "description": "MCP recipe id."
        },
        "stopOnError": {
          "type": "boolean",
          "description": "Stop on first failure."
        },
        "focus_node_id": {
          "type": "string",
          "description": "Work graph node id from agents.work_next packet."
        },
        "continueOnError": {
          "type": "boolean",
          "description": "Continue after failure."
        },
        "idempotency_key": {
          "type": "string",
          "description": "Retry receipt key."
        },
        "semantic_strict": {
          "type": "boolean",
          "description": "Reject batch_semantic_anchor_conflict when anchor set."
        },
        "execution_intent": {
          "enum": [
            "direct",
            "workflow",
            "validate"
          ],
          "type": "string",
          "description": "Batch lane: direct, workflow, or validate."
        },
        "plan_execute_mode": {
          "type": "boolean",
          "description": "Bind plan_execute lane for graph mutations."
        }
      },
      "description": "batch.v1 controls."
    }
  },
  "description": "batch.v1 execute. Direct call: options.execution_intent=direct and no goal. {action, params} is accepted and becomes one step. goal binds a playbook."
}
agentstack.execute output schema
{
  "type": "object",
  "properties": {
    "ok": {
      "type": "boolean",
      "description": "False on total failure."
    },
    "job_id": {
      "type": "string",
      "description": "Poll id."
    },
    "agent_os": {
      "type": "object",
      "description": "Follow next_step."
    },
    "recovery": {
      "type": "object",
      "description": "Next action when blocked."
    },
    "trace_id": {
      "type": "string",
      "description": "Support trace."
    },
    "error_code": {
      "type": "string",
      "description": "Stable failure code."
    },
    "execution_mode": {
      "type": "string",
      "description": "async when queued."
    },
    "partial_success": {
      "type": "boolean",
      "description": "Some steps failed."
    }
  },
  "description": "result.v1 batch envelope."
}
agentstack.execute annotations
{
  "readOnlyHint": false,
  "openWorldHint": true,
  "idempotentHint": false,
  "destructiveHint": false
}

Prompts

72 advertised

create_logic_rule

Advertised by the server during the last successful anonymous check.

Create a backend logic rule from natural language description

description
Natural language description of what the rule should do (required)
project_id
Project ID where the rule will be created (required)
manage_project_users

Advertised by the server during the last successful anonymous check.

Manage users in a project (add, remove, update roles)

project_id
Project ID (required)
action
Action to perform: add, remove, update_role (required)
user_info
User information (email, user_id, role) (required)
create_scheduled_task

Advertised by the server during the last successful anonymous check.

Create a scheduled task that runs on a schedule

name
Task name (required)
cron
Cron expression (e.g., '0 2 * * *' for daily at 2 AM) (required)
command
Command or tool to execute (required)
handle_validation_error

Advertised by the server during the last successful anonymous check.

Guide for handling validation errors

tool_name
Name of the tool that failed (required)
validation_errors
List of validation errors (required)
execute_logic_command

Advertised by the server during the last successful anonymous check.

Execute a logic rule immediately

logic_id
Logic rule ID (UUID) (required)
event_data
Event data to pass to the logic rule
use_processor_in_logic

Advertised by the server during the last successful anonymous check.

Use a processor in a logic rule

processor_name
Name of the processor to use (required)
action
Processor action to execute (required)
parameters
Processor-specific parameters (required)
batch_operations

Advertised by the server during the last successful anonymous check.

Execute multiple commands in batch

commands
Array of command objects to execute (required)
find_entity_by_field

Advertised by the server during the last successful anonymous check.

Find any entity by any field using DNA CRUD

target_entity
Entity type (e.g., 'data_users_user', 'data_projects_8dna') (required)
filters
Filter criteria (any field of the entity) (required)
agentstack_agent_social_participant

Advertised by the server during the last successful anonymous check.

How to participate in AgentSocial chats and graph via MCP social.* (delegated operator token, polling, contracts).

No arguments advertised.

agentstack_system_instructions

Advertised by the server during the last successful anonymous check.

System instructions for AgentStack: when to use which tools, projects, users, and resources (agentstack://me, agentstack://projects/{id}). Use tools for actions; use resources for context.

No arguments advertised.

agentstack_session_setup

Advertised by the server during the last successful anonymous check.

Mandatory session ladder: authenticate, bind context.project_id, then agents.work_next (recipe mcp_work_loop_v1). Catalog search is fallback.

No arguments advertised.

agentstack_stackscript

Advertised by the server during the last successful anonymous check.

StackScript TS subset → SSBC via MCP script.* (not Logic JSON rules).

project_id
Tenant project id (required)
agentstack_work_graph

Advertised by the server during the last successful anonymous check.

Work Graph skill: agents.work_next, plan_get, claim, when blocked, viewer_urls. Same loop for every MCP protocol version.

project_id
Tenant project id (head or organ) (required)
agentstack_closed_loop_autonomy

Advertised by the server during the last successful anonymous check.

Work Graph after session: agents.work_next → follow packet.next_action (often plan_claim → plan_execute when mcp_write). Read keys: work_next without claim. Catalog only if no packet.

project_id
Tenant project id (head or organ) (required)
agentstack_commerce_flows

Advertised by the server during the last successful anonymous check.

Commerce: MCP payments/assets/wallets vs REST marketplace and exchange; canonical action names and typical chains.

use_case
Optional hint: game_economy, saas_billing, marketplace, exchange
agentstack_project_setup

Advertised by the server during the last successful anonymous check.

Step-by-step guide for bootstrapping a complete project with economy, RBAC, logic, scheduler, and API keys

use_case
Optional: game, saas, ecommerce, social, backend_api
agentstack_platform_subscription_grant

Advertised by the server during the last successful anonymous check.

Platform owner: grant personal Premium/VIP to a user or Launch/Business/Scale to a project via buffs

plane
personal | project (required)
tier
premium|vip OR starter|basic|pro|enterprise (required)
entity_id
user_id (personal) or project_id (project) (required)
agentstack_tenant_subscription_monetization

Advertised by the server during the last successful anonymous check.

Tenant project sells subscription SKUs to end users — buff templates, assets, checkout, grant

project_id
Tenant flagship project id (required)
tenant_plan_slug
Tenant SKU slug (e.g. pro_monthly) (required)
agentstack_tenant_buff_grant

Advertised by the server during the last successful anonymous check.

Tenant operator grants subscription buff or inventory to an end user in their project

project_id
Tenant project id (>1) (required)
user_id
Buyer / end-user id (required)
plan_id
Tenant SKU slug (e.g. pro_monthly) (required)
agentstack_project_member_ops

Advertised by the server during the last successful anonymous check.

Add or promote a project member (admin/member/read) — operator and owner paths

project_id
Target project id (required)
user_query
Email or @username to resolve (required)
role
admin | member | read | write
agentstack_generation_canary

Advertised by the server during the last successful anonymous check.

Deprecated alias — use agentstack_tenant_8dna_supply for tenant 8DNA generation supply.

project_id
Project id with sandbox generations (required)
agentstack_showcase_ops

Advertised by the server during the last successful anonymous check.

Ecosystem /showcase gallery catalog on project_id=1 — read, drift, upsert cards, probe health.

No arguments advertised.

agentstack_showcase_vps_ops

Advertised by the server during the last successful anonymous check.

Platform proof showcase (1438–1443): VPS sources sync, hosting.site.build, republish, verify.

No arguments advertised.

agentstack_plan_graph

Advertised by the server during the last successful anonymous check.

Project plan graph (8DNA task_list v2): read, merge, focus run, affinities, F1 promote

No arguments advertised.

agentstack_plan_backlog_ops

Advertised by the server during the last successful anonymous check.

Plan backlog safe cycle: goal-route → plan_process_backlog (preview) → review → plan_apply_proposal (CAS). Telegram/Copilot/MCP parity.

project_id
Project id with orchestrator plan graph (required)
agentstack_inverse_orchestrator_playbook

Advertised by the server during the last successful anonymous check.

Inverse orchestrator for external MCP clients: instruction_packet allowed_actions, advance_run_inverse cursor, wave_xix_atoms on agents.* steps

project_id
Tenant project id for work-graph execution (required)
agentstack_safe_project_cycle

Advertised by the server during the last successful anonymous check.

Master safe cycle: session → preflight → domain mutate in sandbox → verify → diff → gates → promote. All tenant domains via MCP.

project_id
Tenant project id (required)
agentstack_tenant_8dna_supply

Advertised by the server during the last successful anonymous check.

Tenant 8DNA supply chain: auto-fork, diff vs prod, gates, promote, realign, deploy boundary.

project_id
Tenant project id (e.g. flagship Unity tenant) (required)
agentstack_tenant_ops_mcp_first

Advertised by the server during the last successful anonymous check.

MCP-first tenant heal: mentor bots, sandbox verify, promote — no local code unless MCP blocked.

project_id
Tenant project id (required)
agentstack_grants_os_content

Advertised by the server during the last successful anonymous check.

Grant OS content plane: catalog, CRM, earning lanes, anchor/W0 — MCP write paths (ecosystem owner).

program_id
Optional grant program_id to scope earning synergy filter
agentstack_business_organism

Advertised by the server during the last successful anonymous check.

Business organism MCP: head project, organ children, tariff, command snapshot.

head_project_id
Business head project id (optional for bootstrap)
agentstack_tenant_business_assistant

Advertised by the server during the last successful anonymous check.

End-to-end tenant business assistant via MCP recipe (head, CRM organ, operator, Telegram).

No arguments advertised.

agentstack_business_flows

Advertised by the server during the last successful anonymous check.

Quick reference for 12 real business scenarios with tool synergies and cross-flow patterns

scenario
Optional: bootstrap, saas_trial, game_economy, loyalty, ecommerce, automation, content, social, b2b, campaign, usage_billing, matchmaking
agentstack_api_key_safety

Advertised by the server during the last successful anonymous check.

Guide for creating and using API keys safely with service_caps restrictions

No arguments advertised.

agentstack_paradigm_shift

Advertised by the server during the last successful anonymous check.

Paradigm shift: buffs as timers, users as containers, projects as shared state

No arguments advertised.

agentstack_execute_budget

Advertised by the server during the last successful anonymous check.

Sync 60s mcp_batch vs heavy LLM steps; async jobs and discovery.job_status

No arguments advertised.

agentstack_write_modes

Advertised by the server during the last successful anonymous check.

How to edit text and collections: prefer search_replace / unified_diff; merge lists by key; never paste a truncated window as the full body

No arguments advertised.

agentstack_text_editor

Advertised by the server during the last successful anonymous check.

Cross-surface text edit ladder: search_replace and fuzzy unified_diff (same engine as App Studio). Prefer over full-body replace.

No arguments advertised.

agentstack_read_bootstrap

Advertised by the server during the last successful anonymous check.

Read-only MCP bootstrap with continueOnError; omit apikeys.list without api_keys cap

No arguments advertised.

agentstack_social_messenger

Advertised by the server during the last successful anonymous check.

Complete AI guide for using the built-in messenger and social network via MCP (social.chat.*, social.channels.*, social.friends.*, social.public.*)

No arguments advertised.

agentstack_channel_setup

Advertised by the server during the last successful anonymous check.

Unified Channel Plane: enroll delivery sources, category prefs, notifications.deliver, and channels.* MCP tools

No arguments advertised.

agentstack_agents_fleet

Advertised by the server during the last successful anonymous check.

Agents Fleet: configure, enqueue, orchestrate.

project_id
Project id (required)
agentstack_integrations_hub

Advertised by the server during the last successful anonymous check.

Integration Hub: recipes, connections, webhooks, diagnostics.

project_id
Project id (required)
agentstack_crm_workspace

Advertised by the server during the last successful anonymous check.

CRM workspace: contacts, deals, pipeline, import.

project_id
Project id (required)
agentstack_hosted_vertical

Advertised by the server during the last successful anonymous check.

Hosted vertical SaaS: EDITFLOW wire, bootstrap, tenant UI facades.

project_id
Flagship tenant project id (required)
agentstack_key2unity_site

Advertised by the server during the last successful anonymous check.

Key2Unity auth portal /s/1444/key2unity/: persona, bucket, dual bot identity, publish.

project_id
Unity mentor flagship (1444) (required)
agentstack_hosting_storefront

Advertised by the server during the last successful anonymous check.

Hosted shop: create products, list them for sale, publish /s/, take checkout.

project_id
Project id (required)
agentstack_hosting_edit_site

Advertised by the server during the last successful anonymous check.

Change an already published /s/ site with snapshot and release promote (not quickstart).

project_id
Project id (required)
agentstack_hosting_static_landing

Advertised by the server during the last successful anonymous check.

Publish a single HTML landing page to hosted /s/ (agent recipe).

project_id
Project id (required)
agentstack_hosting_workspace_build

Advertised by the server during the last successful anonymous check.

Build static or node_vite site from bucket workspace sources.

project_id
Project id (required)
agentstack_hosting_git_journal

Advertised by the server during the last successful anonymous check.

Per-bucket git journal: init, push dev/main, poll builds, preview and prod URLs.

project_id
Project id (required)
agentstack_deploy_surface

Advertised by the server during the last successful anonymous check.

Resolve deploy surface (site vs app vs shop vs service) before mutating.

project_id
Project id (required)
intent
Operator goal text
agentstack_hosting_react_dist

Advertised by the server during the last successful anonymous check.

Upload pre-built React/Vite dist to /s/ (no npm in MCP).

project_id
Project id (required)
agentstack_hosting_file_edit

Advertised by the server during the last successful anonymous check.

Edit one UTF-8 file in hosted bucket working tree via patch.

project_id
Project id (required)
agentstack_bots_studio

Advertised by the server during the last successful anonymous check.

Bots Fleet studio: templates, brain, handoff, broadcast.

project_id
Project id (required)
agentstack_bot_message_bus

Advertised by the server during the last successful anonymous check.

Agent-to-agent messaging on bot message bus (@handle, webhook push, inbox poll).

project_id
Project id (required)
agentstack_support_staff

Advertised by the server during the last successful anonymous check.

Project support staff inbox and AI binding.

project_id
Project id (required)
agentstack_knowledge_mentor

Advertised by the server during the last successful anonymous check.

Knowledge platform: config leaf, KB ingest, playground (heavy), heal index.

project_id
Project id (required)
agentstack_knowledge_answer_tune

Advertised by the server during the last successful anonymous check.

Tune mentor prompt → playground → promote (UF_KNOWLEDGE_PROMPT_TUNE).

project_id
Project id (required)
agentstack_knowledge_safety_tune

Advertised by the server during the last successful anonymous check.

Safety playbook overlay → safety.evaluate → promote.

project_id
Project id (required)
agentstack_context_lanes

Advertised by the server during the last successful anonymous check.

Pick a context lane, ingest task docs, and tune caps in 8DNA.

project_id
Project id (required)
agentstack_guidance_compass

Advertised by the server during the last successful anonymous check.

Platform Compass playbooks: guidance.list, match_playbook, start_path, complete_step.

project_id
Project id (required)
agentstack_app_studio

Advertised by the server during the last successful anonymous check.

Headless React app studio. Same verbs for an external agent and the project orchestrator.

project_id
Project id (required)
agentstack_project_orchestrator

Advertised by the server during the last successful anonymous check.

Project AI orchestrator — workspace vs messenger vs bots, generation F1, RAG memory.

project_id
Project id (required)
agentstack_rag_memory_orchestrator

Advertised by the server during the last successful anonymous check.

KB vs episodic memory disambiguation for orchestrator agents.

No arguments advertised.

agentstack_project_treasury

Advertised by the server during the last successful anonymous check.

Project wallet treasury vs personal wallets vs AgentNet.

project_id
Project id (required)
agentstack_use_case_game

Advertised by the server during the last successful anonymous check.

Build games with player management, game logic, and monetization

No arguments advertised.

agentstack_use_case_saas

Advertised by the server during the last successful anonymous check.

Build SaaS with trials, subscriptions, and usage tracking

No arguments advertised.

agentstack_use_case_ecommerce

Advertised by the server during the last successful anonymous check.

Build online stores with products, cart, and payments

No arguments advertised.

agentstack_use_case_social

Advertised by the server during the last successful anonymous check.

AgentSocial: messenger, friends graph, PAS, public index via social.* MCP actions

No arguments advertised.

agentstack_use_case_backend_api

Advertised by the server during the last successful anonymous check.

Build REST APIs with CRUD, webhooks, and automation

No arguments advertised.

agentstack_use_case_project_setup

Advertised by the server during the last successful anonymous check.

Bootstrap complete projects: economy, RBAC, logic rules battery, scheduler, FAP, API keys, RAG

No arguments advertised.

Resources

4 advertised

Platform planes

Advertised by the server during the last successful anonymous check.

Project config, storage, hosting, knowledge, CRM, agents, bots, integrations, logic, generation, release

Platform planes URI
agentstack://planes

MIME type: text/plain

Platform planes annotations
{
  "audience": [
    "assistant"
  ],
  "priority": 0.9
}
MCP catalog summary

Advertised by the server during the last successful anonymous check.

Lightweight action totals, catalog etag, registry dimensions

MCP catalog summary URI
agentstack://catalog/summary

MIME type: application/json

MCP catalog summary annotations
{
  "audience": [
    "assistant"
  ],
  "priority": 0.9
}
Session setup ladder

Advertised by the server during the last successful anonymous check.

Mandatory auth → project → context.project_id order (instruction plane SoT)

Session setup ladder URI
agentstack://instructions/session

MIME type: text/markdown

Session setup ladder annotations
{
  "audience": [
    "assistant"
  ],
  "priority": 1
}
Work graph ladder

Advertised by the server during the last successful anonymous check.

Closed-loop plan graph spine — work_next before catalog shopping

Work graph ladder URI
agentstack://instructions/work-graph

MIME type: text/plain

Work graph ladder annotations
{
  "audience": [
    "assistant"
  ],
  "priority": 1
}
Project workspace

Advertised by the server during the last successful anonymous check.

Tenant project JSON config slice (requires membership)

Project workspace URI template
agentstack://projects/{project_id}

MIME type: application/json

Project member

Advertised by the server during the last successful anonymous check.

Project membership row for one user (requires membership)

Project member URI template
agentstack://projects/{project_id}/users/{user_id}

MIME type: application/json

90-day uptime

A lightweight connection check about every six hours, sampled on hosts with many listings. It shows the server answered, not that its tools work.

Too few checks for a percentage yet (1 of 1 answered).

  1. 2026-11-07: 1 of 1 checks answered

30-day availability

Aggregated from read-only remote checks; this is not a tool-execution or security guarantee.

Successful checks
1 of 1 (100%)
Average latency
1272 ms
95th percentile latency
1272 ms

Maintenance history

Public check and source-refresh events. Editorial and ownership records remain private.

  • · probe · passed · 1272 ms · metadata changed · 1 tool(s), 4 resource(s), 2 resource template(s), 72 prompt(s)

Connect from your client

Each snippet registers the same endpoint; pick the client you use.

This server is operated by the named third party. Listing and endpoint checks are not an endorsement or a security audit. Review the operator's permissions, terms, and data handling before connecting. How listings are checked and corrected.

Claude Code

Run this once in a terminal. It registers the remote server. Use --scope project to write .mcp.json for the repo.
Claude Code configuration
claude mcp add --transport http agentstack-mcp https://agentstack.tech/mcp

Codex

Run this once. Codex stores MCP servers in ~/.codex/config.toml (or .codex/config.toml in a trusted project).
Codex configuration
codex mcp add agentstack-mcp --url https://agentstack.tech/mcp

Cursor

Merge this server into .cursor/mcp.json in your project, then enable agentstack-mcp in your MCP settings.
Cursor configuration
{
  "mcpServers": {
    "agentstack-mcp": {
      "url": "https://agentstack.tech/mcp"
    }
  }
}

Cline

Merge this into cline_mcp_settings.json (Cline → MCP Servers → Configure).
Cline configuration
{
  "mcpServers": {
    "agentstack-mcp": {
      "type": "streamableHttp",
      "url": "https://agentstack.tech/mcp",
      "disabled": false
    }
  }
}

VS Code

Merge this server into .vscode/mcp.json. Use MCP: List Servers to start agentstack-mcp and inspect its tools.
VS Code configuration
{
  "servers": {
    "agentstack-mcp": {
      "type": "http",
      "url": "https://agentstack.tech/mcp"
    }
  }
}

Gemini CLI

Run this once. It writes .gemini/settings.json in the current project; pass --scope user for ~/.gemini/settings.json.
Gemini CLI configuration
gemini mcp add --transport http agentstack-mcp https://agentstack.tech/mcp

Claude Desktop

Open Claude Settings > Connectors > Add custom connector, enter this server URL and enable it for the conversation. This is a manual remote-connector handoff, not a desktop JSON configuration.
Claude Desktop configuration
https://agentstack.tech/mcp

Windsurf

Merge this into ~/.codeium/windsurf/mcp_config.json, then refresh MCP in Cascade. Windsurf remote servers use serverUrl, not url.
Windsurf configuration
{
  "mcpServers": {
    "agentstack-mcp": {
      "serverUrl": "https://agentstack.tech/mcp"
    }
  }
}

Let your agent find its tools

Find servers like AgentStack MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.