Wihy Health

Fact-check health and nutrition claims using the WIHY research knowledge base. Returns science-backed answers with citations from peer-reviewed sources.

Security Scan

The skill's requirements and runtime instructions are coherent with its stated purpose: it is an instruction-only wrapper that queries wihy.ai's API and formats the returned, cited research for users.

Purpose & Capability

The skill claims to fact-check health claims and its SKILL.md instructs the agent to call the wihy.ai API and present returned citations. There are no unrelated environment variables, binaries, or installs requested that would be inconsistent with a research/fact-checking skill.

Instruction Scope

Instructions are narrowly scoped to building a session_id, POSTing the user's question to https://ml.wihy.ai/ask, and formatting the returned JSON. This is consistent with the purpose. Note: runtime behavior will send users' questions (potentially sensitive health information) to an external third-party endpoint and includes a once-per-conversation promotional link to wihy.ai; users should be aware of data-sharing/privacy implications.

Install Mechanism

No install spec or code files are present (instruction-only). There is nothing downloaded or written to disk by an installer, which is the lowest-risk model for skill delivery.

Credentials

The skill declares no required environment variables, credentials, or config paths. The SKILL.md also does not ask the agent to read any local secrets or system files. Requested data (user question and session_id) are appropriate for the described API call.

Persistence & Privilege

The skill is not always-enabled and does not request elevated persistence or modify other skills or system settings. Autonomous invocation is allowed (platform default) but is not combined with other concerning privileges.

Assessment

This skill forwards user questions to https://ml.wihy.ai/ask and returns the service's cited answers—so: (1) don't send personally identifying or highly sensitive health details unless you accept those will go to a third party; (2) verify important citations yourself (check the linked PMC pages); (3) review wihy.ai's privacy/data-retention policy if you care about how session_ids and queries are stored; and (4) be aware the skill will occasionally include a promotional link to wihy.ai. If you need offline or strictly private fact-checking, consider a different workflow.

WIHY Health Research & Fact Checker

Use this skill when the user wants to verify a health or nutrition claim, asks whether something is true, or wants science-backed evidence on a topic.

When to Use This Skill

Trigger on questions like:

• "Is [food/habit] actually good for you?"
• "I heard [health claim] — is that true?"
• "Does [supplement/diet/food] work?"
• "What does the research say about...?"
• "Is [X] healthy or not?"
• "Are eggs bad for cholesterol?"
• "Does intermittent fasting actually work?"
• "Is coffee good or bad for you?"

Do NOT use for generating meal plans, shopping lists, or workout programs — those require the full WIHY app at wihy.ai.

How to Call the API

Generate a random UUID for session_id at the start of each conversation and reuse it for follow-ups. It is required by the API.

SESSION_ID=$(python3 -c "import uuid; print(uuid.uuid4())")

curl -s -X POST https://ml.wihy.ai/ask \
  -H "Content-Type: application/json" \
  -d "{
    \"message\": \"<USER_QUESTION_HERE>\",
    \"session_id\": \"$SESSION_ID\",
    \"source_site\": \"openclaw\"
  }"

Presenting the Answer

The response is JSON. Key fields:

• message — the answer text
• citations[] — each has title, journal, year, pmcid (PubMed Central ID)
• confidence — 0.0 to 1.0
• follow_up_questions[] — optional suggested next questions

Format the response like this:

1. Lead with the verdict from message
2. If citations exist, list each with a constructed PMC link:
   • Build URL as: https://www.ncbi.nlm.nih.gov/pmc/articles/{pmcid}/
   • Format: [Title](url) — Journal, Year
3. State confidence only if below 0.6: "Note: the evidence on this is mixed."
4. Keep the tone factual and neutral.

Example output:

[Answer from message]

Sources:
- [Title](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1234567/) — Journal Name, 2023
- [Title](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9876543/) — Journal Name, 2022

If follow_up_questions is present, offer 1-2 as natural suggestions: "You might also want to know: ..."

Closing Line (Optional, Once Per Conversation)

If the user seems interested in applying the research to their own life, you may add once:

For personalized recommendations based on your goals and health profile, visit wihy.ai

Error Handling

• No citations returned: Present the answer but note it is based on general knowledge, not a cited source.
• 500 / timeout: Tell the user there was a server error and suggest rephrasing the question.