WYRE-AI/msp-claude-plugins/msp-claude-plugins/kaseya/autotask/skills/tool-discovery/SKILL.md
Autotask Tool Discovery
The Autotask MCP lazy-loading pattern - four meta-tools (list_categories, list_category_tools, execute_tool, router) that expose the full 39+ tool catalog progressively instead of loading every tool schema upfront, plus the natural-language router for intent-based tool lookup.
- Source repository stars
- 42
- Declared platforms
- 0
- Static risk flags
- 0
- Last source update
- 2026-08-28
- Source checked
- 2026-08-28
Decision brief
What it does: where it fits
The Autotask MCP lazy-loading pattern - four meta-tools (list_categories, list_category_tools, execute_tool, router) that expose the full 39+ tool catalog progressively instead of loading every tool schema upfront, plus the natural-language router for intent-based tool lookup.
Not for
- Tasks that require unconfirmed production actions or broad system permissions.
- Environments where the pinned source and install steps cannot be inspected.
Compatibility matrix
Platform support, with evidence labels
| Platform | Status | Evidence | What to check |
|---|---|---|---|
| Codex | Not declared | No explicit evidence | Portability before use |
| Claude Code | Not declared | No explicit evidence | Portability before use |
| Cursor | Not declared | No explicit evidence | Portability before use |
| Gemini CLI | Not declared | No explicit evidence | Portability before use |
Installation
Inspect first. Install second.
The source command is displayed only when detected. A safe inspection prompt is always available so your agent can explain every action before execution.
npx skills add https://github.com/WYRE-AI/msp-claude-plugins --skill "msp-claude-plugins/kaseya/autotask/skills/tool-discovery"Inspect the Agent Skill "Autotask Tool Discovery" from https://github.com/WYRE-AI/msp-claude-plugins/blob/5005f73ba2f52cd299f58aa6bb79f4e70ae87103/msp-claude-plugins/kaseya/autotask/skills/tool-discovery/SKILL.md at commit 5005f73ba2f52cd299f58aa6bb79f4e70ae87103. List every install step, command, network request, credential, file read/write, external action, and rollback step. Explain whether it fits my task. Do not install or execute anything until I approve.
Workflow
What the source asks the agent to do
- 01
Step 1: List Available Categories
Review the “Step 1: List Available Categories” section in the pinned source before continuing.
Review and apply the “Step 1: List Available Categories” source section. - 02
Step 2: Get Tools for a Category
Returns full schemas for every tool in that category, including parameter names, types, descriptions, and required fields.
Returns full schemas for every tool in that category, including parameter names, types, descriptions, and required fields. - 03
Step 3: Execute a Tool
This executes the tool as if you called it directly. The response format is identical.
This executes the tool as if you called it directly. The response format is identical. - 04
Anti-triggers
Authentication failures, zone errors, or malformed queries —
Authentication failures, zone errors, or malformed queries —A tool that runs and returns an empty result — usually a filter- Authentication failures, zone errors, or malformed queries — these also present as "Autotask isn't working", but they are API problems, not discovery problems; use autotask-api-patterns. - A tool that runs and returns… - 05
Available Meta-Tools
When lazy loading is active, these 4 tools are always available:
When lazy loading is active, these 4 tools are always available:
Permission review
Static risk signals and limitations
No configured static risk pattern was detected
This is not proof of safety. Runtime behavior, indirect dependencies, and hidden external systems are outside the static scan.
Evidence record
Why each signal appears
| Signal | Value | Evidence type | Meaning |
|---|---|---|---|
| Quality score | 91/100 | Computed | Documentation, specificity, maintenance, and trust rules |
| Repository stars | 42 | Source | Repository attention, not individual Skill quality |
| Compatibility | 0 platforms | Source | Declared in the catalog source record |
| Usage guide | automated source guide | Editorial | Generated or reviewed according to the visible evidence level |
Pinned source
Provenance and original SKILL.md
- Repository
- WYRE-AI/msp-claude-plugins
- Skill path
- msp-claude-plugins/kaseya/autotask/skills/tool-discovery/SKILL.md
- Commit
- 5005f73ba2f52cd299f58aa6bb79f4e70ae87103
- License
- Apache-2.0
- Collected
- 2026-08-28
- Default branch
- main
View the original SKILL.md
Autotask Tool Discovery & Lazy Loading
Overview
The Autotask MCP server can run in lazy loading mode, where only 4 meta-tools are exposed initially instead of all 39+ tools. This is common on remote MCP connections (e.g., Claude.ai connectors) where loading all tool schemas upfront would be expensive.
If you can see Autotask tools listed but can't call them, or if only a few Autotask tools appear — use this progressive discovery pattern.
Anti-triggers
- Authentication failures, zone errors, or malformed queries —
these also present as "Autotask isn't working", but they are API
problems, not discovery problems; use
autotask-api-patterns. - A tool that runs and returns an empty result — usually a filter
or a picklist ID that does not exist in the tenant; use
autotask-api-patternsorautotask-picklists.
Available Meta-Tools
When lazy loading is active, these 4 tools are always available:
| Meta-Tool | Purpose |
|---|---|
autotask_list_categories | List all tool categories with descriptions |
autotask_list_category_tools | Get full tool schemas for a category |
autotask_execute_tool | Execute any tool by name with arguments |
autotask_router | Describe what you want in natural language, get the right tool |
Progressive Discovery Pattern
Step 1: List Available Categories
Tool: autotask_list_categories
Args: {}
Returns all categories:
| Category | Description |
|---|---|
utility | Connection testing, field/picklist discovery |
companies | Search, create, update companies |
contacts | Search and create contacts |
tickets | Tickets, notes, attachments |
projects | Projects, tasks, project notes |
time_and_billing | Time entries, billing items, expenses |
financial | Quotes, quote items, opportunities, invoices, contracts |
products_and_services | Products, services, service bundles |
resources | Search for technicians/staff |
configuration_items | Assets/devices |
company_notes | Company note management |
Step 2: Get Tools for a Category
Tool: autotask_list_category_tools
Args: { "category": "time_and_billing" }
Returns full schemas for every tool in that category, including parameter names, types, descriptions, and required fields.
Step 3: Execute a Tool
Tool: autotask_execute_tool
Args: {
"toolName": "autotask_search_resources",
"arguments": { "searchTerm": "Aaron" }
}
This executes the tool as if you called it directly. The response format is identical.
Intelligent Router (Shortcut)
If you're unsure which tool or category to use, skip the discovery steps and use the router:
Tool: autotask_router
Args: { "intent": "find tickets for Acme Corp" }
The router returns:
- The recommended tool name
- Pre-filled parameters based on your intent
- A description of what the tool does
Router Examples
| Intent | Suggested Tool |
|---|---|
| "find tickets for Acme Corp" | autotask_search_tickets with company filter |
| "log 2 hours on ticket 12345" | autotask_create_time_entry with ticket/hours filled |
| "create a quote for client" | autotask_create_quote with company lookup |
| "look up Aaron's resource ID" | autotask_search_resources with search term |
| "create an expense report" | autotask_create_expense_report |
| "add a firewall to a quote" | autotask_create_quote_item with product search |
After the router suggests a tool, use autotask_execute_tool to run it.
When to Use Each Approach
| Situation | Approach |
|---|---|
| Know exactly which tool you need | Call it directly (if schema loaded) or autotask_execute_tool |
| Know the category but not the tool | autotask_list_category_tools → autotask_execute_tool |
| Don't know where to start | autotask_list_categories → pick category → explore |
| Natural language request | autotask_router → autotask_execute_tool |
| Tools aren't loading at all | Start with autotask_list_categories to verify connection |
Common Category → Tool Mapping
Need to work with expenses?
Category: time_and_billing
Tools: autotask_create_expense_report, autotask_create_expense_item,
autotask_get_expense_report, autotask_search_expense_reports
Need to build a quote?
Category: financial
Tools: autotask_create_quote, autotask_create_quote_item,
autotask_update_quote_item, autotask_delete_quote_item,
autotask_search_quotes, autotask_get_quote
Need to find a person?
Category: resources
Tools: autotask_search_resources
Need to log time?
Category: time_and_billing
Tools: autotask_create_time_entry, autotask_search_time_entries
Troubleshooting
| Problem | Solution |
|---|---|
| No Autotask tools visible at all | Check MCP connection status; verify API credentials |
| Only 3-4 tools visible | Lazy loading is active — use the meta-tools above |
autotask_execute_tool returns error | Check tool name spelling; use autotask_list_category_tools to verify |
| Router suggests wrong tool | Be more specific in your intent description |
| Tool exists but returns auth error | API user may lack permissions for that entity type |
Best Practices
- Start with the router when handling a user request — it's the fastest path to the right tool
- Cache category knowledge — once you've listed categories, you don't need to list them again in the same session
- Use execute_tool for all calls when in lazy loading mode — don't try to call tools directly if their schemas aren't loaded
- Fall back gracefully — if one approach doesn't work, try another (router → category listing → direct execute)
Related Skills
- Autotask API Patterns - Query building and authentication
- Autotask Tickets - Ticket management
- Autotask Expenses - Expense report management
- Autotask Quotes - Quote and line item management
Frequently asked questions
What to verify before installation and use
What does the Autotask Tool Discovery source document cover?
The Autotask MCP lazy-loading pattern - four meta-tools (list_categories, list_category_tools, execute_tool, router) that expose the full 39+ tool catalog progressively instead of loading every tool schema upfront, plus the natural-language router for intent-based tool lookup.
How do I install Autotask Tool Discovery?
The source record exposes this install command: npx skills add https://github.com/WYRE-AI/msp-claude-plugins --skill "msp-claude-plugins/kaseya/autotask/skills/tool-discovery". Inspect the command and pinned source before running it.