research[B]-idea

user_requirements.md priority: All user constraints recorded in docs/user_requirements.md (direction preferences, implementation requirements, document format, etc.) take precedence over any default instruction in this skill. Always read that file before generating any output to ensure compliance with confirmed user constraints.

Phase B: Idea Deepening

Starting from a confirmed research direction and RQs, deepen the "research question" into an "implementable method" through three layers of confirmation. Produces docs/idea_report.md Part 2, then moves to Phase C.

Prerequisite: docs/idea_report.md exists and contains Part 1 (Phase A complete).

Workflow Overview & Outputs

ResearchPilot-Skills splits a complete academic research project into seven independent stage skills. The current skill is one link in that chain.

Seven-Stage Chain

Project Directory Structure

docs/
  idea_report.md        # Research report in three parts:
                        #   Part 1: Motivation, RQs, Key Works  (Phase A)
                        #   Part 2: Introduction, Related Works, Method  (Phase B)
                        #   Part 3: Datasets, Experiment Design, Resource Estimate  (Phase C)
  implementation.md     # Coding guide: file- and function-level implementation plan  (Phase D)
  dev_log.md            # Dev log: progress, decisions, "How to Run" section  (Phase E)
  user_requirements.md  # User constraints: collected and maintained by Claude
  papers/               # Downloaded paper PDFs or abstract TXTs
  manuscripts/          # Paper drafts, each revision archived separately
                        #   (e.g. v1.0-initial-draft.md, v1.1-revision.md)

code/
  src/                  # Core model and training code
  scripts/              # Run scripts (train.sh, evaluate.sh, ablation.sh)
  configs/              # Hyperparameter config files
  baselines/            # Baseline model implementations
  notebooks/            # Visualization notebooks; paper figure/table generation
  data/                 # Datasets (gitignored)
  results/              # Experiment results (gitignored)
  logs/                 # Training logs (gitignored)
  README.md             # Environment setup and run commands
  requirements.txt      # Dependencies (library names only, no torch family)

---

Command

/research[B]-idea

An optional natural-language instruction may follow the command. The AI will treat it as an additional constraint or supplement for this invocation.

---

Phase B Flow Overview

B-0 Compile Part 1 (directly from Phase A confirmed content — no regeneration)
B-1 Confirm technical framework (Layer 1)
B-2 Confirm detailed pipeline (Layer 2 — plain language, step-by-step)
    → Write Part 2: Related Works + Method
B-3 Verify cited content (locate supporting sentences in PDFs)
B-4 Polish Introduction (Layer 3)
B-5 Ask user whether to proceed to Phase C

Full step-by-step instructions: references/phase-B.md.

---

Confirmation Card

Every output in this phase begins with a confirmed-content card:

━━━━━━━━━━ Confirmed Content ━━━━━━━━━━
Research direction: {confirmed}
Primary RQ: {confirmed}
Secondary RQs: {confirmed}
Direction constraints: {user constraints}
RQ constraints: {user constraints}
Reference papers: {user-specified}
Technical framework: {confirmed framework, one sentence}
Pipeline: {confirmed pipeline, one sentence}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Only output confirmed, non-empty fields. Omit empty or unconfirmed fields entirely.

---

Non-Negotiable Constraints

1. Do not advance to Phase C without explicit user confirmation.
2. Never fabricate citations. Unverifiable content gets ⚠️ [low confidence: reason] and is logged in the pending-verification list.
3. Before writing or revising any idea, read existing literature in docs/papers/ extensively (see references/phase-B.md, Literature Reading Principle).
4. The pipeline output must use plain language — "Step 1…Step 2…" — not formulas.
5. Rules in references/template-flexibility.md take precedence over any specific template instruction.

---

On Phase Completion

After Part 2 is fully confirmed:

Phase B complete. idea_report.md Part 2 has been generated.

After Part 2 is fully confirmed:

Phase B complete. idea_report.md Part 2 has been generated.

→ Use /research[C]-experiment to enter the Experiment Design phase.

---

Reference Files

• Detailed flow: references/phase-B.md
• Document format specs: references/document-formats.md
• Template flexibility rules: references/template-flexibility.md