scout

name

scout

description

Quick preliminary research on a supplement or question. Searches for promising evidence that might warrant full deep research. Outputs a structured scouting brief to wiki/scouting/ — kept separate from the main wiki until promoted through deep research and ingest. Use when: "is X worth researching deeper?", "what's the preliminary evidence for X?", "$scout taurine cognitive", "should I research quercetin?". Do NOT use for comprehensive research (use foundry-research deep-research instead) or for wiki ingestion (use wiki-ingest).

scout

You are conducting a quick preliminary investigation of a supplement or health question. The goal is to determine whether something is promising enough to warrant a full deep-research session, not to produce definitive answers.

Output Silo

All scouting output goes to wiki/scouting/ — it is not part of the main wiki. This keeps preliminary findings separate from ingested, vetted knowledge.

When to Use

• "Is X worth researching deeper?"
• "What's the preliminary evidence for X?"
• "$scout " or "$scout "
• "Should I research quercetin?"
• Quick gap-filling: "what does the literature say about X interaction?"

When NOT to Use

• Full supplement investigation → use foundry-research deep-research
• Ingesting into the wiki → use wiki-ingest
• Answering from existing wiki knowledge → search the wiki first

Steps

1. Read context

• Read wiki/handoff.md
• Read purpose.md
• Check if the supplement/topic already exists in the wiki (search wiki/entities/ and wiki/scouting/)
• If a scouting brief already exists, offer to update it rather than create a duplicate

2. Understand the scope

Parse the user's request into:

• Subject: supplement name or topic
• Angle: specific outcome, pathway, or question (if specified; otherwise "general promise")
• Context: why they're asking (e.g., considering adding to stack, saw a claim, gap from previous research)

If the scope is unclear, ask one focused question — do not present a menu.

3. Search (5-10 min equivalent)

Execute a targeted search following the methodology in wiki/docs/research-methodology.md:

Academic-first search:

• Search PubMed/Google Scholar for recent reviews (2024+) on the subject
• Search for the highest-evidence studies (meta-analyses > RCTs > observational)
• Note study types and sample sizes

Supplementary search:

• Check Examine.com for a summary if available
• Check regulatory status (FDA, EMA) if relevant
• Quick safety check — known adverse effects, drug interactions

Minimum source targets:

• 1-2 systematic reviews or meta-analyses (if they exist)
• 2-3 primary studies (RCTs preferred)
• 1 safety/interaction source

If fewer than 5 sources are findable, note the scarcity — that itself is a finding.

4. Assess promise

Rate the preliminary evidence:

Rating	Meaning	Next step
High promise	Multiple RCTs with positive outcomes, clear mechanism, good safety	Recommend deep research + eventual ingest
Moderate promise	Some human evidence, plausible mechanism, limited but positive data	Recommend deep research to clarify
Low promise	Mostly animal/in-vitro, weak human data, or mixed results	Flag as low priority; deep research only if specifically interested
Dead end	Negative RCTs, safety concerns, or no credible human evidence	Do not recommend further research
Insufficient data	Too few sources to assess	Recommend deep research if the question matters, otherwise deprioritize

5. Write the scouting brief

Create a file in wiki/scouting/ using this structure:

---
type: entity
entity_type: supplement
aliases: []
sources: []
created: "YYYY-MM-DD"
updated: "YYYY-MM-DD"
status: preliminary
tags:
  - supplement/<name>
  - scouting
---

> [!tldr]
> One-sentence verdict on promise level.

## Promise Rating

**[High / Moderate / Low / Dead end / Insufficient data]**

## Why Investigated

<!-- Brief context on why this was scouted -->

## What We Found

### Evidence Snapshot

| Evidence Level | Count | Key Finding |
|----------------|-------|-------------|
| Meta-analysis/systematic review | | |
| Clinical endpoint RCT | | |
| Biomarker RCT | | |
| Observational | | |
| Animal/in-vitro | | |

### Top 3 Sources

1. **Source 1** — Key finding. [Tier] (full text / abstract only)
2. **Source 2** — Key finding. [Tier] (full text / abstract only)
3. **Source 3** — Key finding. [Tier] (full text / abstract only)

### Mechanism Summary

<!-- 2-3 sentences on known or proposed mechanism -->

### Safety Signal

<!-- Any red flags? Green flags? Unknown? -->

## What's Missing

> [!gap]
> - [ ] What questions remain unanswered
> - [ ] What evidence level is needed to confirm promise

## Recommendation

<!-- Clear recommendation: deep research? deprioritize? specific angle to investigate? -->

## Search Methodology

- Databases searched: [e.g., Google Scholar, PubMed]
- Sources screened: [number]
- Sources read: [number full text, number abstract only]
- Date of search: YYYY-MM-DD
- Limitations: [any gaps in the search]

6. Present to user

Show the user:

1. Promise rating with brief justification
2. Top finding (one sentence)
3. Recommendation (deep research? deprioritize?)
4. Path to the scouting brief file

7. Update handoff

Append a note to wiki/handoff.md about the scouting session. Log in wiki/log.md.

Promoting to Deep Research

If the scouting brief rates "High promise" or "Moderate promise" and the user wants to proceed:

1. Use the scouting brief as input to frame the deep-research query
2. The deep-research report supersedes the scouting brief
3. After deep research completes, the scouting brief gets a superseded-by frontmatter field pointing to the research report
4. The wiki-ingest skill then ingests from the research report as normal

Scouting Brief Lifecycle

wiki/scouting/<Supplement>.md
  → rated High/Moderate → deep research → wiki-ingest → scouting brief marked superseded
  → rated Low → kept in scouting/ as reference, not promoted
  → rated Dead end → kept in scouting/ with clear "dead end" status

Scouting briefs are never directly ingested into the main wiki. They inform deep research queries.