/research

name: research
description: 'Run a multi-source research session searching GitHub, HN, Lobsters, Reddit, arXiv, and Semantic Scholar. Use for multi-channel topic surveys.'
usage: >-
  /tome:research "<topic>" [--format report|brief|transcript]
  [--resume] [--list] [--domain <override>]

Research Command

Start an autonomous multi-source research session.

Usage

# Full research session
/tome:research "async python patterns"

# Choose output format
/tome:research "cache eviction policies" --format brief

# Resume most recent session
/tome:research --resume

# List saved sessions
/tome:research --list

# Override domain classification
/tome:research "raft consensus" --domain architecture

What Happens

1. Topic is classified into a domain (algorithm, ui-ux, architecture, etc.) 2. Research channels are selected based on domain 3. Parallel agents search each channel autonomously 4. Findings are merged, deduplicated, and ranked 5. A formatted report is generated and saved 6. You can refine results with `/tome:dig`

Execution

Invoke the research skill:

Skill(tome:research)

Pass the topic, format flag, and any overrides.

Output

Reports are saved to `docs/research/{session}-{topic}.md`.

See Also

• `/tome:dig`: refine results interactively
• `/tome:cite`: generate bibliography
• `/tome:export`: export to memory-palace format