菜单
Health domain memory classification — taxonomy, permanence levels, and example fact patterns for storing health knowledge with entity resolution
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Tzeusy/butlers --skill memory-taxonomy命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Guide for discovering, analyzing, and pruning the Butlers test suite. Use when working on test condensation beads (Phase 1 epic bu-rhztl and Phase 2 epic bu-hg8rl both CLOSED; Phase 3 maintenance cycle underway 2026-06-21), assessing test bloat, identifying pruning targets, or rewriting tests to be contract-driven. Triggers on test reduction, test pruning, test consolidation, or condensation tasks for this project. Also use when a fresh session needs to assess test health, create new condensation beads, or resume in-progress condensation work.
Generate a weekly home energy digest with trends, top consumers, and recommendations.
Orchestrate a UX redesign of a Butlers dashboard page (or sub-page set) using /project-direction as the spec+beads engine, with redesign-specific upfront phases for vision capture, asset ingestion, impact analysis, backend-contract derivation, LLM-cost feasibility, manifesto/identity preservation, and a th-design design-bar audit. The binding design language is the Dispatch spec (openspec/specs/dashboard-design-language/spec.md); bundles live under pr/overview/ and resolve via references/bundle-registry.md. Use when asked to redesign a dashboard page, with or without a Claude Design bundle. Triggers on "redesign the X page", "plan the Y redesign", "integrate the redesign bundle", "what would it take to ship the SLUG redesign", "design language integration for AREA".
正在显示 SKILL.md
基于 SOC 职业分类
| name | memory-taxonomy |
| description | Health domain memory classification — taxonomy, permanence levels, and example fact patterns for storing health knowledge with entity resolution |
| version | 2.0.0 |
| tools_required | ["memory_store_fact","memory_recall","memory_search","memory_entity_resolve","memory_entity_create"] |
Use this skill when storing health facts to memory. It defines the health domain taxonomy for subjects, predicates, permanence levels, and tagging — ensuring consistent, retrievable health knowledge over time.
Only load this skill when you are actively extracting and storing facts from a user message. It is not needed for every session.
For the full entity resolution protocol — including the resolve-or-create transitory pattern,
disambiguation policy, and idempotency handling — see the butler-memory shared skill.
Every fact about a healthcare provider, pharmacy, clinic, or other external health entity
MUST be anchored to a resolved entity via entity_id. Never call memory_store_fact with
only a raw subject string for external entities.
User-related health facts (conditions, medications, allergies, symptoms) are about the user
themselves — use the sender's entity_id from the REQUEST CONTEXT preamble.
When calling or creating a transitory entity, infer the correct from context:
memory_entity_resolveentity_type| Health entity | entity_type |
|---|---|
| Healthcare provider (doctor, specialist, therapist) | person |
| Clinic, hospital, medical practice, imaging lab | organization |
| Pharmacy (Walgreens, CVS, local pharmacy) | organization |
| Insurance provider | organization |
| User's own health facts (conditions, meds, allergies) | person — use sender entity_id from preamble |
| Medication or condition name as topic | (no entity required — anchor fact to user's entity_id) |
When a healthcare provider or pharmacy is first mentioned and not in the entity graph, create a transitory entity:
# "Dr. Chen is my primary care physician"
# Step 1: resolve the provider
candidates = memory_entity_resolve(
name="Dr. Chen",
entity_type="person",
context_hints={"topic": "primary care physician"}
)
# → zero candidates: create transitory entity
try:
result = memory_entity_create(
canonical_name="Dr. Chen",
entity_type="person",
metadata={
"unidentified": True,
"source": "fact_storage",
"source_butler": "health",
"source_scope": "health"
}
)
provider_entity_id = result["entity_id"]
except ValueError:
# Entity already exists (concurrent creation) — resolve to get entity_id
candidates = memory_entity_resolve(name="Dr. Chen", entity_type="person")
provider_entity_id = candidates[0]["entity_id"]
# Step 2: store edge-fact from user to provider
memory_store_fact(
subject="user",
predicate="doctor_primary_care",
content="Dr. Chen",
entity_id="<sender_entity_id>", # user's entity from preamble
object_entity_id=provider_entity_id, # edge-fact: user → provider
permanence="stable",
importance=7.0,
tags=["healthcare-provider"]
)
The provider entity appears in the dashboard "Unidentified Entities" section for the owner to confirm, merge, or delete. Never fall back to bare string subjects for external entities.
entity_id from REQUEST CONTEXT preamble; subject label is the user's name or "user"person entity; use provider name as subject labelorganization entity; use institution name as subject labelsubject label, but anchor with the user's entity_id| Predicate | Meaning |
|---|---|
medication | Current medication with dosage |
medication_frequency | How often taken |
dosage | Amount per dose |
condition_status | "active", "managed", or "resolved" |
symptom_pattern | Recurring symptoms or triggers |
symptom_trigger | What causes or worsens a symptom |
measurement_baseline | Typical or target measurement values |
dietary_restriction | Food allergies or restrictions |
exercise_routine | Regular physical activity |
doctor_name | Healthcare provider name (fact on provider entity) |
doctor_primary_care | User's primary care physician — edge-fact (user → provider entity) |
pharmacy | Preferred pharmacy — edge-fact (user → pharmacy entity) |
allergy | Medication or substance allergies |
| Level | When to use |
|---|---|
stable | Chronic conditions, long-term medications, allergies — things unlikely to change |
standard | Current medications, active symptoms, dietary patterns — current state |
volatile | Acute symptoms, temporary conditions, one-time measurements |
Use tags to enable cross-cutting queries. Common health tags:
chronic, acute, medication, condition, tracking, goal, allergy, critical, blood-pressure, diabetes, symptom, pattern, healthcare-provider, dosage-change, sensitive, private
# From: "Started taking Lisinopril 10mg daily for blood pressure"
# entity_id comes from sender preamble (user's entity)
memory_store_fact(
subject="Lisinopril",
predicate="medication",
content="10mg daily for blood pressure management",
entity_id="<sender_entity_id>",
permanence="standard",
importance=8.0,
tags=["medication", "blood-pressure"]
)
# From: "I'm allergic to penicillin"
memory_store_fact(
subject="user",
predicate="allergy",
content="allergic to penicillin",
entity_id="<sender_entity_id>",
permanence="stable",
importance=9.0,
tags=["allergy", "critical"]
)
# From: "Headaches usually happen when I don't drink enough water"
memory_store_fact(
subject="headaches",
predicate="symptom_trigger",
content="triggered by dehydration",
entity_id="<sender_entity_id>",
permanence="standard",
importance=6.0,
tags=["symptom", "pattern"]
)
# From: "Dr. Chen is my primary care physician"
# Step 1: resolve or create provider entity (see Resolve-or-Create above)
# Step 2: store edge-fact from user to provider
memory_store_fact(
subject="user",
predicate="doctor_primary_care",
content="Dr. Chen",
entity_id="<sender_entity_id>",
object_entity_id="<dr_chen_entity_id>", # transitory entity for Dr. Chen
permanence="stable",
importance=7.0,
tags=["healthcare-provider"]
)
# From: "I use Walgreens on Main Street for prescriptions"
# Step 1: resolve or create pharmacy entity
candidates = memory_entity_resolve(name="Walgreens", entity_type="organization",
context_hints={"topic": "pharmacy"})
# → use existing or create transitory (see butler-memory resolve-or-create protocol)
# Step 2: store edge-fact
memory_store_fact(
subject="user",
predicate="pharmacy",
content="Walgreens on Main Street",
entity_id="<sender_entity_id>",
object_entity_id="<walgreens_entity_id>",
permanence="stable",
importance=6.0,
tags=["healthcare-provider", "pharmacy"]
)
stable; acute symptoms are volatilesensitive or private tags for personal health information