khamel83/research

/research — Background Research via Argus

Run deep research in a background sub-agent. Argus handles search provider selection and fallback.

Usage

/research [topic]

Process

1. Ask 2-3 clarifying questions (goal, depth, audience)

2. Check Argus availability:

   python -c "from core.search.argus_client import is_available; print(is_available())" 2>/dev/null
   

   Or simply:

   curl -s http://localhost:8005/api/health >/dev/null 2>&1 && echo "argus: yes" || echo "argus: no"
   

3. Mode 1: Argus (Primary) Spawn background sub-agent:

   Task:
     subagent_type: general-purpose
     description: "Argus research: [topic]"
     run_in_background: true
     prompt: |
       Use Argus to research: [topic]
       
       1. Query Argus in research mode:
          curl -s -X POST http://localhost:8005/api/search \
            -d '{"query": "[research query]", "mode": "research"}'
       
       2. For key results, extract full content:
          curl -s -X POST http://localhost:8005/api/extract \
            -d '{"url": "RESULT_URL"}'
       
       3. Write findings to: docs/research/[slug]/research.md
       4. Structure: Executive Summary → Detailed Findings → Sources
   

4. Mode 2: Gemini CLI (Fallback) If Argus is unavailable:

   Task:
     subagent_type: general-purpose
     description: "Gemini research: [topic]"
     run_in_background: true
     prompt: |
       Run Gemini CLI for research:
       gemini --yolo "[research prompt]"
       
       Save output to: docs/research/[slug]/research.md
   

5. Present findings when agent completes

6. Save to docs/research/[slug]/research.md

Search Config

Argus modes (from config/search.yaml):

• discovery: broad, multiple sources
• precision: targeted, high relevance
• research: comprehensive, all providers
• cheap: SearXNG only, cost-sensitive

Notes

• Research takes 3-8 minutes in background
• Check docs/research/ for past research
• Argus automatically handles provider fallback, ranking, and dedup