Research Review

If the project is outside this domain, adapt to the closest equivalent senior field-journal reviewer.

Context: $ARGUMENTS

Prerequisites

• Use spawn_agent and send_input when the user has explicitly allowed delegation or subagents.
• If delegation is not allowed, run the same review loop locally and preserve the same deliverable structure.

Workflow

Step 1: Gather Research Context

Before calling the external reviewer, compile a comprehensive briefing:

1. Read project narrative documents such as STORY.md, README.md, or paper drafts
2. Read notes for key findings and experiment history
3. Identify core claims, methodology, key results, and known weaknesses
4. If the topic is multi-agent formation control, explicitly summarize:
   • agent dynamics assumptions
   • communication graph assumptions
   • disturbance, delay, or obstacle assumptions
   • scalability evidence
   • theorem vs simulation vs real-system evidence boundaries

Step 2: Initial Review (Round 1)

Send a detailed prompt with xhigh reasoning:

spawn_agent:
  model: gpt-5.4
  reasoning_effort: xhigh
  message: |
    [Full research context + specific questions]

    Please act as a senior professor in multi-agent systems, formation control,
    cooperative robotics, and distributed control, with extensive IEEE journal
    reviewing experience.

    Identify:
    1. Logical gaps or unjustified claims
    2. Missing experiments that would materially strengthen the paper
    3. Weaknesses in framing, novelty, or evidence sufficiency
    4. Whether the contribution is strong enough for an IEEE field journal
    5. For control papers specifically: weak assumptions, missing robustness checks,
       missing scalability analysis, or theorem-to-simulation mismatches

    Please be brutally honest and prioritize the minimum fixes with the highest acceptance lift.

Step 3: Iterative Dialogue (Rounds 2-N)

Use send_input with the returned agent id to continue the conversation.

For each round:

1. Respond to criticisms with evidence or counterarguments
2. Ask targeted follow-ups on the most actionable points
3. Request concrete deliverables such as experiment designs, paper outlines, or claims matrices

Key follow-up patterns:

• "If we reframe X as Y, does that change your assessment?"
• "What is the minimum experiment to satisfy concern Z?"
• "Please design the minimal additional experiment package with the highest acceptance lift per GPU week."
• "Please write a mock IEEE journal review with scores."
• "Give me a results-to-claims matrix for possible outcomes."

Step 4: Convergence

Stop iterating when:

• both sides agree on the core claims and their evidence requirements
• a concrete experiment plan is established
• the narrative structure is settled

Step 5: Document Everything

Save the full interaction and conclusions to a review document in the project root:

• round-by-round summary of criticisms and responses
• final consensus on claims, narrative, and experiments
• claims matrix
• prioritized TODO list with estimated costs
• paper outline if discussed

Update project memory or notes with key review conclusions.

Key Rules

• ALWAYS use reasoning_effort: xhigh for reviews
• send comprehensive context in Round 1 because the external model cannot read local files
• be honest about weaknesses
• push back on criticisms you disagree with, but accept valid ones
• focus on actionable feedback
• document the agent id for potential future resumption
• the review document should be self-contained

Prompt Templates

For initial review

"I'm going to present a complete research project for your critical review. Please act as a senior professor in multi-agent systems, formation control, cooperative robotics, and distributed control with extensive IEEE journal reviewing experience..."

For experiment design

"Please design the minimal additional experiment package that gives the highest acceptance lift per GPU week. Be specific about datasets, agent counts, graph conditions, disturbances, delays, and baselines."

For paper structure

"Please turn this into a concrete IEEE-style paper outline with section-by-section claims and figure plan."

For claims matrix

"Please give me a results-to-claims matrix: what claim is allowed under each possible outcome of experiments X and Y?"

For mock review

"Please write a mock IEEE journal review with Summary, Strengths, Weaknesses, Questions for Authors, Score, Confidence, and What Would Move Toward Accept."