K-Dense-AI/scientific-agent-skills / skills/treatment-plans
treatment-plans
VerifiedFormats and structurally validates local treatment-plan documentation after clinical decisions have already been supplied and verified by authorized licensed professionals. Use for source traceability, clinician-authored intervention records, goals and checkpoints, shared-decision records, reconciliation handoffs, and release gates—not for clinical decision-making.
references/safety_scope.md
View on GitHub# Safety, Scope, and Routing
Last reviewed: **2026-10-01**
## Purpose
This skill is a transcription, formatting, provenance, and process-validation aid. It accepts clinical decisions only after an authorized licensed professional has made and verified them in a current local source.
The intended use excludes clinical decision support, prescribing, medication checking, triage, and patient education. This description is not a regulatory classification of any deployment; qualified institutional reviewers must determine applicable requirements.
## Prohibited functions
Do not:
- identify or infer a diagnosis, differential, severity, stage, risk class, or eligibility;
- propose, compare, rank, select, substitute, or optimize a treatment;
- generate medication names, doses, routes, frequencies, durations, start dates, stop dates, hold criteria, titration steps, or taper schedules;
- judge an interaction, allergy, contraindication, precaution, organ-function issue, pregnancy issue, or formulary suitability;
- generate a monitoring parameter, target, threshold, interval, follow-up frequency, or escalation criterion;
- interpret a symptom, test, image, score, trend, medication list, adverse event, or patient preference;
- determine urgency, triage disposition, emergency status, prognosis, expected response, or likely outcome;
- create patient-specific instructions, education, warning signs, crisis plans, or emergency actions;
- recommend a specialist, setting, service level, procedure, device, or referral;
- certify compliance, clinical completeness, standard of care, informed consent, capacity, or professional scope.
Formatting a supplied decision does not validate it. A citation does not make a decision current or applicable.
## Request handling
Proceed only when the request is equivalent to:
- "Place these already signed clinician decisions into the generic records."
- "Check whether this local JSON package has the required fields."
- "Verify that every record points to a verified source fact."
- "List checkpoints on the exact dates already provided."
- "Identify missing acknowledgments or sign-off fields without suggesting clinical content."
Stop and route when the request asks:
- "What should the plan be?"
- "What treatment, medication, dose, or schedule is best?"
- "Should this be started, stopped, held, resumed, increased, reduced, or tapered?"
- "Are these medicines safe together or contraindicated?"
- "Is this urgent, an emergency, or likely to worsen?"
- "What should the patient do now?"
Do not soften a prohibited request into a recommendation-shaped template. Do not ask another skill, model, search tool, or API to make the decision.
## Missing or conflicting content
When a required clinical field is missing:
1. Leave it empty or mark the record pending.
2. Record a nonclinical blocker code and field path.
3. Route it to the responsible authorized professional.
4. Do not infer a value from neighboring records, standard practice, prior examples, a product label, or a guideline.
When sources conflict, record the conflict without deciding which source controls. The authorized local team must reconcile it in an approved system.
## Emergency and escalation boundary
Use this exact process statement:
> If a concern may be urgent or emergent, stop this documentation workflow and use the institution's current clinical escalation or emergency process; this package does not determine urgency or provide emergency instructions.
The package records only the local process reference and responsible role. It must not include generated symptom thresholds, emergency numbers, destinations, or action steps. Current institution-approved material may be linked by a local reference after authorized review.
## Accountable roles
At minimum, identify:
- clinical owner — owns clinical decisions and conflict resolution;
- authorized licensed verifier — compares transcribed content with current sources;
- medication-reconciliation owner — performs reconciliation in approved systems;
- privacy reviewer — reviews patient-derived data handling and de-identification claims;
- records/governance owner — controls retention, access, versioning, and release;
- handoff sender and recipient — own transfer and acknowledgment;
- local reporting owner — determines whether and where an event must be reported.
One person may hold multiple roles only if local policy permits it. A script can verify that roles are named; it cannot verify competence, licensure, authority, independence, or completion.
## Release boundary
The documentation package remains visibly marked:
> **DRAFT — NOT MEDICAL ADVICE — DOCUMENTATION-ONLY — AUTHORIZED CLINICIAN SIGN-OFF REQUIRED**
Release means only that the package may enter the authorized documentation handoff named in the manifest. It does not turn the package into stand-alone medical advice or authorize an agent to implement care.
Block release when any of these remain:
- unverified facts or interventions;
- missing source links;
- unresolved discrepancies not routed to an owner;
- absent shared-decision documentation when required by the local workflow;
- incomplete reconciliation;
- missing privacy/process attestations;
- missing current local policy or labeling verification;
- missing handoff recipient or acknowledgment;
- unsigned clinician attestation;
- inconsistent subject reference, classification, status, or dates.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file). | Stars: 198k | Updated | Claude CodeCodex | Unverified | Install |
| doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |