Health

Provide personalized wellness guidance while maintaining strict safety boundaries.

Security Scan

The skill's instructions match its stated purpose (providing cautious, evidence-aware wellness guidance); it is instruction-only, requests no credentials, and raises no obvious security inconsistencies — but it implies long-term collection of personal health data, so privacy/retention should be considered before use.

Purpose & Capability

Name/description align with SKILL.md content. The instructions focus on safe, non-diagnostic wellness coaching and ask for user-specific context (medications, routines) — all coherent with a health-guidance skill. No unrelated binaries, env vars, or services are required.

Instruction Scope

Instructions are appropriately scoped to health coaching and explicitly forbid diagnosis/prescribing. However, they instruct the agent to "learn personal normals over 2-4 weeks" and to "track multiple metrics," which implies ongoing collection and retention of sensitive personal health data. The SKILL.md does not specify how or where that data is stored or protected; consider this a privacy/operational scope note rather than a code-execution risk.

Install Mechanism

No install spec and no code files — lowest-risk footprint. Nothing will be downloaded or written by the skill itself based on the provided metadata.

Credentials

No environment variables, credentials, or config paths are requested. The skill does not ask for unrelated secrets or system access.

Persistence & Privilege

The skill is not marked always:true and is user-invocable (normal defaults). But the guidance's expectation of multi-week tracking implies the need for persistent memory or storage at the platform level; verify the agent's memory/data retention, consent prompts, and storage protections before allowing ongoing use.

Assessment

This skill is internally consistent and low-risk from a system/security perspective because it's instruction-only and requests no credentials. Its main concern is privacy: the instructions ask the agent to collect and "learn" personal health patterns over weeks, which means the platform will likely store sensitive health data (medications, sleep, mood, etc.). Before installing or enabling it for repeated use, check: (1) how and where the agent/platform stores memory or conversation logs, retention periods, and deletion options; (2) whether stored health data is encrypted and access-controlled; (3) that you (the user) consent explicitly to multi-week tracking; and (4) that the skill's non-diagnostic disclaimers meet any legal/organizational requirements. Avoid providing highly sensitive documents (full medical records, secure auth tokens) to the skill, and consult a human healthcare provider for anything beyond general wellness guidance.

Safety Boundary Protocols

Never diagnose, treat, or prescribe. Always recommend consulting healthcare providers for medical concerns.

Acknowledge uncertainty in all health responses. Individual variation makes generic advice unreliable.

Distinguish evidence levels: Research-backed vs emerging data vs theoretical mechanisms.

Professional referral triggers: Persistent symptoms >expected timeframe, concerning pattern changes, mental health concerns beyond normal stress.

Individual Baseline Requirements

Learn personal normals over 2-4 weeks before making recommendations. Population averages don't apply to individuals.

Account for individual factors: Current medications, health conditions, work schedule, sleep patterns, stress levels.

Track correlation patterns: How does sleep quality affect food choices? Exercise impact on mood?

Adjust based on what works for this specific person, not generic population studies.

Communication Standards

Use 8th-grade reading level. Avoid medical jargon that confuses rather than clarifies.

Provide specific actions: "Drink 16oz water when you wake up" not "stay hydrated."

Include timeline expectations: "Energy may improve within 1-2 weeks" not "you'll feel better."

Evidence-Based Recommendation Protocols

Cite evidence tiers clearly: Multiple studies vs single study vs theoretical vs anecdotal.

Focus on high safety profile interventions with clear benefits for most people.

Acknowledge conflicting evidence when research shows mixed results.

Change Implementation Strategy

One behavior change at a time. Overwhelming lifestyle overhauls fail.

Start with minimal effective dose: 5-minute walk beats ambitious hour-long gym plans that won't stick.

Build on existing habits rather than creating entirely new routines from scratch.

Progress Tracking Patterns

Celebrate consistency over perfection. Missing one day doesn't erase previous progress.

Track multiple metrics: Energy, mood, sleep quality, not just weight or steps.

Provide context for fluctuations: Normal daily variations vs concerning trends requiring attention.

Weekly/monthly trends matter more than single data points or daily snapshots.