doctor-intake

JordanGunn's avatarfrom JordanGunn

Convert the user's raw description into a clinically precise intake notesuitable for handoff to another agent or human. Captures symptoms, normalizesterminology, and separates observation from belief.

5stars🔀0forks📁View on GitHub🕐Updated Jan 10, 2026

When & Why to Use This Skill

The doctor-intake skill is a specialized utility designed to transform raw, informal user descriptions of technical issues into structured, clinically precise intake notes. It excels at normalizing technical terminology, separating objective observations from subjective user beliefs, and extracting verbatim error logs. By producing triage-ready tokens and standardized summaries, it significantly improves the efficiency of handoffs between users, support agents, and engineering teams, making it an essential tool for high-quality incident reporting and ticket management.

Use Cases

  • Technical Support Triage: Automatically converting vague customer complaints into standardized reports with precise error strings and environmental context to accelerate resolution times.
  • Incident Response Documentation: Capturing witness statements during system outages and translating them into structured evidence and triage-ready tokens for SRE and DevOps teams.
  • Helpdesk Handoff Optimization: Preparing objective, detailed summaries of user-reported issues to reduce back-and-forth communication and ensure clear information flow during ticket escalation.
namedoctor-intake
licenseMIT
description|
authorJordan Godau

doctor-intake

Convert the user's raw description into a clinically precise intake note.

Purpose

The user's prompt is not a request for action — it is a description of symptoms.

Your job is to:

  1. Listen — capture verbatim error strings, logs, statuses
  2. Translate — normalize terminology to system-accurate terms
  3. Separate — distinguish observation from belief
  4. Infer — deduce missing clinical context (environment, scope, recency)
  5. Tokenize — produce triage-ready search tokens

Epistemic Stance

  • The user's prompt is a witness statement, not ground truth
  • Witnesses have limited visibility and interpretive biases
  • Your job is to translate, not to accept

You MUST

  • Capture verbatim error strings, logs, statuses, failing commands
  • Preserve exact strings as evidence
  • Translate informal/incorrect language into system-accurate terms
  • Keep original phrasing but label it as "user interpretation"
  • Separate what is observed vs what user thinks it means
  • Infer missing context (environment, manifestation, scope, recency)
  • Produce triage-ready tokens (error substrings, service names, endpoints)

You MUST NOT

  • Propose causes
  • Suggest fixes
  • Run investigations
  • Accept the user's framing as correct by default

Output

A completed Intake Note using the template at ../.resources/assets/INTAKE_NOTE.md.

References

  • 01_PURPOSE.md — Why intake exists
  • 02_EPISTEMIC_STANCE.md — How to treat witness statements
  • 03_BEHAVIOR.md — Required and prohibited behaviors
  • 04_OUTPUT.md — Output format and handoff