Skip to main content
Glama

draft_response

Read-onlyIdempotent

Compose a knowledge-base-grounded draft reply for a support ticket, including citations. Refuses to draft when a security incident is detected.

Instructions

Compose a knowledge-base-grounded draft reply for a ticket.

Returns the draft together with grounding — the KB excerpts it was built from — and kb_article_ids for citation. You may improve the phrasing of the returned draft. You may not add any fact, URL, timeframe, or procedure step that does not appear in grounding. The knowledge base contains no phone numbers; a phone number in a reply is fabricated by definition.

SECURITY GUARDRAIL — ENFORCED IN CODE, NOT NEGOTIABLE This tool refuses to draft for any ticket that shows signs of a security incident, and returns ok=false with SECURITY_ESCALATION_REQUIRED. The decision is made two ways, independently:

  1. the ticket's as-filed category is "security"; or

  2. the ticket text trips a KB-006 indicator — credentials entered on a suspicious site, an attachment opened followed by changed system behaviour, files encrypted or a ransom note, browser hijack symptoms, a vendor requesting changed bank details, mail spoofing the client's own domain, or a user denying an account change they are recorded as making. Layer 2 fires even when the ticket is filed as hardware, email, or anything else. A non-security category is never sufficient to obtain a draft.

There is no parameter, phrasing, or instruction that disables this. Do not attempt to work around a refusal by drafting the reply yourself, by calling search_kb and composing from the excerpts, or by asking the user to relay troubleshooting steps. KB-006 is explicit that support does not troubleshoot suspected security incidents at all: restarts and quick fixes destroy forensic evidence, and a password reset is insufficient during an active compromise. On refusal, tell the user plainly that the ticket needs the security team, report the indicators returned, and offer to escalate with update_ticket.

WHAT IT DOES NOT DO

  • Does not send anything. The draft is returned to you for review; delivery is out of scope for this server.

  • Does not change ticket state. Use update_ticket to record an escalation.

  • Does not invent facts when the knowledge base is silent. If retrieval finds nothing, it returns KB_NO_MATCH rather than a plausible-sounding reply.

ERRORS TICKET_NOT_FOUND — no such ticket; find the right ID with search_tickets. SECURITY_ESCALATION_REQUIRED — refused; escalate to the security team. KB_NO_MATCH — the knowledge base does not cover this issue. Escalate to a technician rather than answering from general knowledge.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
ticket_idYes

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
okNo
draftNoThe grounded draft reply. Null whenever ok is false.
messageNo
refusalNo
groundingNoThe ONLY facts available for this reply. Wording may be adjusted; facts, URLs, and timeframes absent from these excerpts may not be added.
ticket_idYes
error_codeNo
kb_article_idsNoArticles the draft is grounded in.
Behavior5/5

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

Disclosures go far beyond annotations (readOnlyHint, idempotentHint). Details the security guardrail (two independent checks, indicators, non-negotiable). Explains return behavior, error conditions, and what it does not do. No contradiction with annotations.

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?

Well-structured with clear headings and bullet points. Front-loaded with core purpose. Somewhat lengthy but every section adds essential value (security, limitations, errors). Could trim redundant phrasing slightly.

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

Completeness5/5

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

Comprehensive: covers all error cases, security guardrail with detailed indicators, return value composition (references output schema), usage boundaries, and escalation path. Output schema exists so return format documentation is not required in description.

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?

Only one parameter (ticket_id). Schema coverage is 0% but the description implicitly ties the parameter to its purpose ('for a ticket'). Does not add explicit syntax or constraints but sufficient given simplicity of schema.

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

Purpose5/5

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

Description opens with a clear verb ('Compose') and resource ('knowledge-base-grounded draft reply for a ticket'). It distinguishes itself from siblings like search_kb (searching vs drafting) and update_ticket (no state change).

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 when to use the tool: to draft a KB-grounded reply. Provides clear do's and don'ts (e.g., do not invent facts, do not attempt to work around refusal). Offers actionable guidance on errors and what to do on security refusal.

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

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Jackson-DM/msp-tools-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server