Interrogate Your Research Question With Socratic Dialogue4 skills
Guided exploration of research questions, methodology, and evidence foundations through intent-based Socratic questioning.
Research & Writing Blueprint
10-stage pipeline for academic research, writing, peer review, and revision with mandatory integrity checks and honesty safeguards.
Run this command to deploy the blueprint to your environment.
Academic research requires coordination across multiple expertise areas (ideation, writing, reviewing, revision) with high stakes for accuracy. You need multiple perspectives on your work—a fresh reviewer, a methodologist, a critical devil's advocate—but managing these stages manually is time-consuming. AI can accelerate this process, but introducing AI into academic research creates new risks: hallucinated citations, frame-locked reasoning, and subtle sycophancy that compounds errors across stages. You need a structured pipeline that enforces honesty primitives at each stage.
You start by describing your research question, and the system guides you through a Socratic dialogue to clarify your methodology. Once approved, you move to writing: the system generates an outline, helps you build arguments with proper citations, and catches plagiarism and citation errors before peer review. A seven-agent review team (EIC, three domain reviewers, a methodologist, a devil's advocate) evaluates your paper and produces a decision. If revisions are needed, you engage in Socratic coaching with the EIC, revise, and undergo re-review. Integrity verification runs at two mandatory gates (pre-review and post-revision) to catch hallucinations and verify claims against retrieved sources. Once approved, you finalize your paper in your chosen format (Markdown, DOCX, PDF, or LaTeX) and receive a process summary with collaboration quality feedback.
Guided exploration of research questions, methodology, and evidence foundations through intent-based Socratic questioning.
Outline-first writing pipeline with style calibration, argument mapping, citation compliance, and anti-leakage safeguards.
7-mode failure checklist and sampling-based claim verification to catch hallucinations, fabrications, and methodological flaws before external review.
Seven-agent review panel (EIC, 3 domain reviewers, methodologist, devil's advocate, synthesizer) with sprint contracts and 0-100 quality rubrics.
EIC-led Socratic dialogue to guide revision strategy, then AI-assisted drafting with change tracking and delta reporting.
Narrow re-review team (3 agents) verifies your revisions meet reviewer requirements, updates decision, and checks for regressions.
Zero-tolerance deep re-check of all claims after revision to ensure no regressions and all modifications are sound.
Format selection, venue-specific AI disclosure, output rendering (Markdown, DOCX, LaTeX, PDF), and optional cross-session resume.
Socratic Guided Research
Intent-based dialogue that distinguishes exploratory thinking from goal-oriented planning. Guides you through RQ clarification without premature closure.
Multi-Perspective Review
Seven-agent panel (EIC + 3 domain reviewers + methodologist + devil's advocate + synthesizer) with 0-100 quality rubrics and sprint contract pre-commitment.
Mandatory Integrity Gates
Two non-skippable gates (pre-review and post-revision) run AI failure mode checklists: hallucinated citations, fabricated methods, frame-lock, shortcut reliance, and claim-faithfulness audits.
Anti-Sycophancy Protocols
Devil's advocate must score rebuttals before conceding. Collaboration depth observer tracks delegation vs. cognitive vigilance. Reading-check probe verifies you engaged with cited sources.
Cross-Session Resume & Passports
Material Passport (JSON/YAML) preserves research artifacts, provenance, and audit trails. Resume from any checkpoint in a fresh session without losing context.
Socratic Guided Research
Intent-based dialogue that distinguishes exploratory thinking from goal-oriented planning. Guides you through RQ clarification without premature closure.
Multi-Perspective Review
Seven-agent panel (EIC + 3 domain reviewers + methodologist + devil's advocate + synthesizer) with 0-100 quality rubrics and sprint contract pre-commitment.
Mandatory Integrity Gates
Two non-skippable gates (pre-review and post-revision) run AI failure mode checklists: hallucinated citations, fabricated methods, frame-lock, shortcut reliance, and claim-faithfulness audits.
Anti-Sycophancy Protocols
Devil's advocate must score rebuttals before conceding. Collaboration depth observer tracks delegation vs. cognitive vigilance. Reading-check probe verifies you engaged with cited sources.
Cross-Session Resume & Passports
Material Passport (JSON/YAML) preserves research artifacts, provenance, and audit trails. Resume from any checkpoint in a fresh session without losing context.